From jonstanley at gmail.com Thu Oct 1 03:36:55 2009 From: jonstanley at gmail.com (Jon Stanley) Date: Wed, 30 Sep 2009 23:36:55 -0400 Subject: [fedora-virt] merging lists? Message-ID: I'm working on migrating Fedora lists from RHT infrastructure to Fedora infrastructure in the coming little bit. The question that I had was does fedora-xen and fedora-virt both need to exist? I'm subscribed to both, and certainly the conversations on each are different, but could they be merged? As of several months ago, the subscriber count of fedora-xen was WAY higher than that of fedora-virt, this doesn't seem to make much sense since Xen is not The Way Forward(TM), at least from what I can see. I guess the question is could we merge the two lists, and have one less to migrate, or are both really needed? From clalance at redhat.com Thu Oct 1 06:26:49 2009 From: clalance at redhat.com (Chris Lalancette) Date: Thu, 01 Oct 2009 08:26:49 +0200 Subject: [fedora-virt] merging lists? In-Reply-To: References: Message-ID: <4AC44BA9.2010104@redhat.com> Jon Stanley wrote: > I'm working on migrating Fedora lists from RHT infrastructure to > Fedora infrastructure in the coming little bit. The question that I > had was does fedora-xen and fedora-virt both need to exist? I'm > subscribed to both, and certainly the conversations on each are > different, but could they be merged? > > As of several months ago, the subscriber count of fedora-xen was WAY > higher than that of fedora-virt, this doesn't seem to make much sense > since Xen is not The Way Forward(TM), at least from what I can see. > > I guess the question is could we merge the two lists, and have one > less to migrate, or are both really needed? I would vote for merging them. I don't see a real need to have the two separate lists; we can discuss both Xen and KVM on the fedora-virt list with no problem. I think the fact that fedora-xen has so many subscribers is historical, since that was the first virtualization technology. A polite note to fedora-xen saying that the list is going away, and to re-subscribe to fedora-virt, would probably work. It's worked for other lists we've renamed/combined. -- Chris Lalancette From clalance at redhat.com Thu Oct 1 06:34:56 2009 From: clalance at redhat.com (Chris Lalancette) Date: Thu, 01 Oct 2009 08:34:56 +0200 Subject: [fedora-virt] using kqemu with fedora-virt-preview repo In-Reply-To: <561c252c0909301654r76bde27cve99147120b69bc25@mail.gmail.com> References: <561c252c0909301654r76bde27cve99147120b69bc25@mail.gmail.com> Message-ID: <4AC44D90.80007@redhat.com> Gianluca Cecchi wrote: > I have a powerful server but without VT capabilities, where I would like > to use virt-manager tools as of fedora-virt-preview repo and speed up > somehow things using kqemu. > It seems that kqemu support is not compiled in for this repo, correct? > Instead, as the kqemu kernel module is in rpmfusion, I presume that the > stock f11 qemu has this support built in.... > Am I wrong? I do think you are wrong, as you found out below; we have kqemu specifically disabled. kqemu is going away, and in fact has been ripped out of upstream qemu, because it only kind-of sort-of works some of the time. > Any problems to compile kqemu support ? > > On another server, tried to rebuild qemu-0.11.0-2.fc11.src.rpm (from > fedora-virt) after installing kqemu from rpmfusion. > But while if I take the source I get kqemu as enabled in configure > script, in rpmbuild I don't get so. > I notice that inside the qemu tar.gz as provided by the source.rpm there > are three lines below in the top configure script: > > if [ "$cpu" = "i386" -o "$cpu" = "x86_64" ] ; then > kqemu="yes" > audio_possible_drivers="$audio_possible_drivers fmod" > kvm="yes" > kqemu="no" > > So I commented out the kqemu="no" line... and built the package (not > touching the other two configure scripts that seems related to kvm...) > After installing, if I run qem-kvm or qemu without kqemu all is ok. > Instead, running > [root at tekkaman bin]# ./qemu --enable-kqemu > I get > Not enough memory (requested_size = 16777216, max memory = 146800640) > Aborted > > Any hints? > For Kqemu now gplv2, are there any barriers to have it included in > fedora directly? Yes, it's completely unmaintained, not upstream, and never will be. Sorry, the way of the future is KVM (and possibly Xen dom0 support, if and when it makes it into upstream Linux). -- Chris Lalancette From markmc at redhat.com Thu Oct 1 07:19:38 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Thu, 01 Oct 2009 08:19:38 +0100 Subject: [fedora-virt] merging lists? In-Reply-To: References: Message-ID: <1254381578.2920.21.camel@blaa> (cross-posting to fedora-xen since it affects folks only subscribed there) On Wed, 2009-09-30 at 23:36 -0400, Jon Stanley wrote: > I'm working on migrating Fedora lists from RHT infrastructure to > Fedora infrastructure in the coming little bit. The question that I > had was does fedora-xen and fedora-virt both need to exist? I'm > subscribed to both, and certainly the conversations on each are > different, but could they be merged? > > As of several months ago, the subscriber count of fedora-xen was WAY > higher than that of fedora-virt, this doesn't seem to make much sense > since Xen is not The Way Forward(TM), at least from what I can see. > > I guess the question is could we merge the two lists, and have one > less to migrate, or are both really needed? We discussed this earlier this month: http://www.redhat.com/archives/fedora-xen/2009-September/thread.html#00021 and decided to keep fedora-xen as a place mainly to discuss the Dom0 pv_ops work, since that currently isn't in Fedora and there's been a lot of traffic relating to it. However, if we're moving to a new server and starting with a clean slate, I don't think it would make sense to re-create the fedora-xen list. We'll want to close it down sooner or later, so we may as well use that event to do it. Cheers, Mark. From rampal.anuj at gmail.com Thu Oct 1 08:30:49 2009 From: rampal.anuj at gmail.com (anuj rampal) Date: Thu, 1 Oct 2009 14:00:49 +0530 Subject: [fedora-virt] virt-viewer windows port In-Reply-To: <20090930162125.GD32184@amd.home.annexia.org> References: <20090923145706.GA31875@amd.home.annexia.org> <20090924081844.GA32184@amd.home.annexia.org> <20090924132443.GB32184@amd.home.annexia.org> <20090925120254.GC32184@amd.home.annexia.org> <20090930162125.GD32184@amd.home.annexia.org> Message-ID: > i tried "./configure --sysconfdir=C:\pki\" > but this didnot work... When you say this "did not work", how did it fail? The clue is usually in the error message. when i did --sysconfdir="C:\pki\" I expected that i have keep my certificates under this directory(correct me if i'm wrong). But it was still taking the certificates from "Z:\usr\i686-pc-mingw\sys-confic\mingw\etc\pki" (in Windows). -------------------------- Regarding virsh: *virsh on linux mahine:* URI tried from a remote linux mahine: virsh -c qemu+tcp://FC11-KVM/session virsh -c qemu://FC11-KVM/session virsh -c qemu+ssh://FC11-KVM/session virsh -c qemu+unix://FC11-KVM/session all of these work fine * virsh on windows:* on windows the supported URIs are: virsh.exe -c qemu+tcp://FC11-KVM/session virsh.exe -c qemu://FC11-KVM/session URI "qemu+tcp://FC11-KVM/session" works fine. It gets connected and i can call all the functions. I have also written a small code on C#.net and i use this URI (qemu+tcp://FC11-KVM/session) to connect to libvirt and call the functions and it works fine without any problem. Now the problem again comes with URI : virsh.exe -c qemu://FC11-KVM/session when i rum me code using the above URI this is what the error message that it gives: Cannot access CA certificate '/usr/i686-pc-mingw32/sys-root/mingw/etc/pki/CA/cacert.pem' : errno=2 So basically i need to change the the path it is looking in..... On Wed, Sep 30, 2009 at 9:51 PM, Richard W.M. Jones wrote: > On Wed, Sep 30, 2009 at 09:02:59PM +0530, anuj rampal wrote: > > i tried "./configure --sysconfdir=C:\pki\" > > but this didnot work... > > When you say this "did not work", how did it fail? The clue is > usually in the error message. > > In any case, you almost certainly can't just write C:\pki\ because \ > is an escape character. Maybe you can double them, or use forward > slashes instead. > > > LIBVIRT_DEBUG=1 virt-viewer -c qemu://FC11-KVM/session 2 > > It's really not helpful to use virt-viewer to test. Test using a > simple tool like virsh. Get that working first. Once you understand > what the problem was, *only then* try virt-viewer. > > > Do we have to do any configuration on libvirtd to make virt-viewer to > work > > over tcp and tls because it works fine from a remote linux machine over > ssh. > > You can also disable encryption in libvirtd: > > http://libvirt.org/remote.html#Remote_libvirtd_configuration > > or read the libvirtd.conf file. > > Rich. > > -- > Richard Jones, Emerging Technologies, Red Hat > http://et.redhat.com/~rjones > Read my programming blog: http://rwmj.wordpress.com > Fedora now supports 75 OCaml packages (the OPEN alternative to F#) > http://cocan.org/getting_started_with_ocaml_on_red_hat_and_fedora > -------------- next part -------------- An HTML attachment was scrubbed... URL: From rjones at redhat.com Thu Oct 1 08:46:46 2009 From: rjones at redhat.com (Richard W.M. Jones) Date: Thu, 1 Oct 2009 09:46:46 +0100 Subject: [fedora-virt] virt-viewer windows port In-Reply-To: References: <20090923145706.GA31875@amd.home.annexia.org> <20090924081844.GA32184@amd.home.annexia.org> <20090924132443.GB32184@amd.home.annexia.org> <20090925120254.GC32184@amd.home.annexia.org> <20090930162125.GD32184@amd.home.annexia.org> Message-ID: <20091001084646.GB32442@amd.home.annexia.org> On Thu, Oct 01, 2009 at 02:00:49PM +0530, anuj rampal wrote: > > i tried "./configure --sysconfdir=C:\pki\" > > but this didnot work... > > When you say this "did not work", how did it fail? The clue is > usually in the error message. > > when i did --sysconfdir="C:\pki\" I expected that i have keep my > certificates under this directory(correct me if i'm wrong). > But it was still taking the certificates from > "Z:\usr\i686-pc-mingw\sys-confic\mingw\etc\pki" (in Windows). But what were the messages in the configure step? Rich. -- Richard Jones, Emerging Technologies, Red Hat http://et.redhat.com/~rjones virt-df lists disk usage of guests without needing to install any software inside the virtual machine. Supports Linux and Windows. http://et.redhat.com/~rjones/virt-df/ From gianluca.cecchi at gmail.com Thu Oct 1 09:15:36 2009 From: gianluca.cecchi at gmail.com (Gianluca Cecchi) Date: Thu, 1 Oct 2009 11:15:36 +0200 Subject: [fedora-virt] using kqemu with fedora-virt-preview repo In-Reply-To: <4AC44D90.80007@redhat.com> References: <561c252c0909301654r76bde27cve99147120b69bc25@mail.gmail.com> <4AC44D90.80007@redhat.com> Message-ID: <561c252c0910010215r4aae319dud0ffe1103879e0aa@mail.gmail.com> On Thu, Oct 1, 2009 at 8:34 AM, Chris Lalancette wrote: > > I do think you are wrong, as you found out below; we have kqemu > specifically > disabled. kqemu is going away, and in fact has been ripped out of upstream > qemu, because it only kind-of sort-of works some of the time. > Ok, I will keep this information and in fact found references that also in f11 stock, even if one installs kqemu package from rpmfusion, he/she receives the error ... What exactly do you mean with "ripped out of upstream"? That I cannot use at all the tgz provided inside the src.rpm? At least I would like to try something, as I read happy Ubuntu users regarding kqemu and recent qemu and kernel versions.... and with very good performance in guests too... Having a BL25p G1 server with two Dual core Amd 275 cpus and getting so slow VMs is a pity... in my opinion > > > Any problems to compile kqemu support ? > > > > > Yes, it's completely unmaintained, not upstream, and never will be. Sorry, > the > way of the future is KVM (and possibly Xen dom0 support, if and when it > makes it > into upstream Linux). > OK. But perhaps if one could at least test... Thanks, Gianluca -------------- next part -------------- An HTML attachment was scrubbed... URL: From andres at verot.com Thu Oct 1 17:37:05 2009 From: andres at verot.com (=?ISO-8859-1?Q?Andr=E9s_Garc=EDa?=) Date: Thu, 01 Oct 2009 19:37:05 +0200 Subject: [fedora-virt] parameters Message-ID: <4AC4E8C1.9020707@verot.com> Hi, I want to set up wo virtual WinXP machines to act as licence managers for a couple of CAD/CAM programs. With the first one, everything went well, but when I got the second one to work, the virtual machines got confused, as if they didn't know which was their dongle. I think it is because in their definition files there is something like: ... ... The vendor and product ids are the same in both cases. I saw in the docs that I can't change it to:
'bus' and 'slot' I can get with 'lsusb', but what does 'function' mean? Thanks, Andr?s From markmc at redhat.com Thu Oct 1 20:59:01 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Thu, 01 Oct 2009 21:59:01 +0100 Subject: [fedora-virt] virt-preview updates In-Reply-To: <1250703309.17874.46.camel@blaa> References: <1244034684.5001.134.camel@blaa> <1244046483.5001.180.camel@blaa> <1244197459.27876.0.camel@blaa> <1246303956.11688.148.camel@blaa> <1246628572.13604.2.camel@blaa> <1246636154.13604.8.camel@blaa> <1247763853.3038.105.camel@blaa> <1248769817.3089.20.camel@blaa> <1248776732.3089.45.camel@blaa> <1248859649.3021.28.camel@blaa> <20090729093637.GC10723@salstar.sk> <1248868438.3021.31.camel@blaa> <1248873640.3021.34.camel@blaa> <1248879462.3021.39.camel@blaa> <1248882336.3021.40.camel@blaa> <1248974623.3275.4.camel@blaa> <1249057022.3320.35.camel@blaa> <1249400197.3212.73.camel@blaa> <1249573056.3721.62.camel@blaa> <1249639097.3260.5.camel@blaa> <1249900612.8784.79.camel@blaa> <1250700893.17874.44.camel@blaa> <1250703309.17874.46.camel@blaa> Message-ID: <1254430742.2920.48.camel@blaa> The latest updates available from http://markmc.fedorapeople.org/virt-preview == libvirt == * Thu Oct 1 2009 Mark McLoughlin - 0.7.1-8 - Disable sound backend, even when selinux is disabled (#524499) - Re-label qcow2 backing files (#497131) == python-virtinst == * Thu Sep 24 2009 Cole Robinson - 0.500.0-4.fc12 - Don't use usermode net for non-root qemu:///system via virt-install - Fix cdrom installs where the iso is a storage volume (bz 524109) - Fix path permissions for kernel/initrd download location (bz 523960) * Wed Sep 16 2009 Cole Robinson - 0.500.0-3.fc12 - Don't generate bogus disk driver XML. - Add '--disk format=' for specifying format (qcow2, ...) when provisioning - Add Fedora12 to os dictionary == virt-manager == * Tue Sep 29 2009 Cole Robinson - 0.8.0-6.fc12 - Fix VCPU hotplug - Remove access to outdated docs (bz 522823, bz 524805) - Update VM state text in manager view (bz 526182) - Update translations (bz 493795) * Thu Sep 24 2009 Cole Robinson - 0.8.0-5.fc12 - Refresh host disk space in create wizard (bz 502777) - Offer to fix disk permission issues (bz 517379) * Thu Sep 17 2009 Cole Robinson - 0.8.0-4.fc12 - Don't close libvirt connection for non-fatal errors (bz 522168) - Manager UI tweaks - Generate better errors if disk/net stats polling fails Cheers, Mark. == About The Virtualization Preview Repository == The virt-preview repository offers Fedora 11 users the opportunity to help out with testing the rawhide virtualization packages, without having to run all of rawhide. Bugs should be reported to http://bugzilla.redhat.com/ using 'rawhide' as the distribution version. Note clearly in the bug report that you are running these packages on Fedora 11. From bill at bfccomputing.com Fri Oct 2 02:49:43 2009 From: bill at bfccomputing.com (Bill McGonigle) Date: Thu, 01 Oct 2009 22:49:43 -0400 Subject: [fedora-virt] merging lists? In-Reply-To: References: Message-ID: <4AC56A47.3000005@bfccomputing.com> On 09/30/2009 11:36 PM, Jon Stanley wrote: > I guess the question is could we merge the two lists, and have one > less to migrate, or are both really needed? Just one data point - my procmail has already merged the lists. ;) -Bill -- Bill McGonigle, Owner BFC Computing, LLC http://bfccomputing.com/ Telephone: +1.603.448.4440 Email, IM, VOIP: bill at bfccomputing.com VCard: http://bfccomputing.com/vcard/bill.vcf Social networks: bill_mcgonigle/bill.mcgonigle From jonstanley at gmail.com Fri Oct 2 03:13:03 2009 From: jonstanley at gmail.com (Jon Stanley) Date: Thu, 1 Oct 2009 23:13:03 -0400 Subject: [fedora-virt] merging lists? In-Reply-To: <4AC56A47.3000005@bfccomputing.com> References: <4AC56A47.3000005@bfccomputing.com> Message-ID: On Thu, Oct 1, 2009 at 10:49 PM, Bill McGonigle wrote: > Just one data point - my procmail has already merged the lists. ;) :) I suspect that many others have as well, which was the point of the thread. I do recall the discussion on fedora-xen, but the pvops dom0 stuff, while not being non-existant traffic, certainly doesn't overwhelm my mailbox to the point that I've created any sort of filter for it. From gianluca.cecchi at gmail.com Fri Oct 2 13:36:49 2009 From: gianluca.cecchi at gmail.com (Gianluca Cecchi) Date: Fri, 2 Oct 2009 15:36:49 +0200 Subject: [fedora-virt] virt-preview updates Message-ID: <561c252c0910020636s77442eedy919ba63a4b44b4f5@mail.gmail.com> On Thu, 01 Oct 2009 21:59:01 +0100 Mark McLoughlin wrote: > == virt-manager == > > * Tue Sep 29 2009 Cole Robinson - 0.8.0-6.fc12 > - Fix VCPU hotplug Hello, any info on about what would have been fixed? And also provided functionality itself.... In my env with 0.8.0-6 and f11 with virt-preview it seems I cannot hotplug/unplug at all a cpu.... In details window of a rh 5.3 vm with 2 vCPUs, I cannot raise the number (current is equal to maximum), while I can lower but I get that the effect will be visible at next reboot..... Or do I have to create a new VM with this peculiar version of virt-manager to see different behaviour? Gianluca -------------- next part -------------- An HTML attachment was scrubbed... URL: From crobinso at redhat.com Fri Oct 2 13:46:42 2009 From: crobinso at redhat.com (Cole Robinson) Date: Fri, 02 Oct 2009 09:46:42 -0400 Subject: [fedora-virt] virt-preview updates In-Reply-To: <561c252c0910020636s77442eedy919ba63a4b44b4f5@mail.gmail.com> References: <561c252c0910020636s77442eedy919ba63a4b44b4f5@mail.gmail.com> Message-ID: <4AC60442.60607@redhat.com> On 10/02/2009 09:36 AM, Gianluca Cecchi wrote: > On Thu, 01 Oct 2009 21:59:01 +0100 Mark McLoughlin wrote: >> == virt-manager == >> >> * Tue Sep 29 2009 Cole Robinson - 0.8.0-6.fc12 >> - Fix VCPU hotplug > > Hello, > any info on about what would have been fixed? And also provided > functionality itself.... > In my env with 0.8.0-6 and f11 with virt-preview it seems I cannot > hotplug/unplug at all a cpu.... > In details window of a rh 5.3 vm with 2 vCPUs, I cannot raise the > number (current is equal to maximum), while I can lower but I get that > the effect will be visible at next reboot..... > > Or do I have to create a new VM with this peculiar version of > virt-manager to see different behaviour? > There was basically a syntax error behind the scenes when trying to use the libvirt 'SetVcpus' call. Libvirt actually doesn't even support cpu hotplug for qemu/kvm yet so it wouldn't work anyways. I should have written a better changelog entry. - Cole From markmc at redhat.com Fri Oct 2 17:48:26 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Fri, 02 Oct 2009 18:48:26 +0100 Subject: [fedora-virt] Fedora virt status Message-ID: <1254505706.4346.18.camel@blaa> Fedora 12 ========= The Fedora 12 release is drawing closer and closer. Beta is on its way out the door and release candidate composes will begin in less than four weeks time. Helping Out =========== Here are three ways you could help out with getting F12 into great shape: 1) Test, test, test and file bugs. The more stuff you break, the more will get fixed. Maybe you could even write some test cases; for examples, just take a look at: https://fedoraproject.org/wiki/Test_Day:2009-09-17_Virtualization https://fedoraproject.org/wiki/Test_Day:2009-05-07_Virtualization 2) Follow the action in bugzilla. See this wiki page for how to sign up to receive all bugzilla emails relating to Fedora virtualization: https://fedoraproject.org/wiki/Virtualization_bugs 3) Take a look at some existing bugs and see if you can help out getting them closed. A good place to start is the F12 target list: https://bugzilla.redhat.com/showdependencytree.cgi?id=F12VirtTarget&hide_resolved=1 But be warned, you may quickly find yourself becoming a package co-maintainer! :-) Test Day ======== Earlier this month, a dedicated bunch of folks got together on irc for the day to see what they could break: https://fedoraproject.org/wiki/Test_Day:2009-09-17_Virtualization To give you an idea of the success of the day, here's a selection of the bugs reported: https://bugzilla.redhat.com/523953 (mbanas) libvirtd segfault with NIC hot-unplug https://bugzilla.redhat.com/524012 (vbenes) libvirt cannot hot-unplug devices which were not previously hot-plugged https://bugzilla.redhat.com/524022 (vbenes) qemu gets confused with NIC hotplug when no model is specified https://bugzilla.redhat.com/523914 (jdenemar) Mouse does not move in PV Xen guest under RHEL-5.4 https://bugzilla.redhat.com/523941 (jdenemar) kernel 2.6.31-1[24].fc12 doesn't boot in xen PV guest https://bugzilla.redhat.com/523971 (jdenemar) xm save hangs with kernel-2.6.31-14.fc12 running as a PV guest under RHEL-5.4 https://bugzilla.redhat.com/523939 (mrezanin) Save on restored machine failed https://bugzilla.redhat.com/523623 (drjones) grub.conf needs console=hvc0 in kernel command-line when installed as DomU https://bugzilla.redhat.com/524039 (drjones) block device cannot be detached from DomU https://bugzilla.redhat.com/524052 (drjones) Boot hang when installing HVM DomU https://bugzilla.redhat.com/524035 (vbenes) libvirt should support USB device detach https://bugzilla.redhat.com/524045 (vbenes) accessing non existing usb device cause guest to stop https://bugzilla.redhat.com/523158 (santwana) libvirtd segault during virsh restore https://bugzilla.redhat.com/523956 (clalance) Starting libvirtd by hand causes denials https://bugzilla.redhat.com/523970 (jstodola) virsh help output not sorted https://bugzilla.redhat.com/523993 (clalance) KVM Live migration failure with SELinux Enforcing https://bugzilla.redhat.com/524047 (gianluca) virt-viewer outputs nothing if connecting to a non existent VM https://bugzilla.redhat.com/522683 (plambert) USB devices do not mount and are not seen by 12-alpha KVM https://bugzilla.redhat.com/523948 (phan) VM cannot boot from the disk converted by qemu-img https://bugzilla.redhat.com/524229 (santwana) Local migration of kvm guest fails in Fedora12 Alpha https://bugzilla.redhat.com/524268 (santwana) KVM guest fails to start up after virt-snapshot in Fedora12 https://bugzilla.redhat.com/524033 (jstodala) libvirt should check the permissions on all paths before starting qemu https://bugzilla.redhat.com/523960 (mnowak) virtinst saves images where qemu can't access them by default https://bugzilla.redhat.com/517379 (adamw) virt-manager should warn if guest images will are not readable by qemu https://bugzilla.redhat.com/524083 (emcnabb) virt-manager storage "Allocation" field can be set higher than "Max Capacity" https://bugzilla.redhat.com/524085 (emcnabb) Help button on virt-manager "New Storage Volume" page broken https://bugzilla.redhat.com/524109 (lmr) virt-manager: Fails to install guest using ISO file - internal error unable to start guest: qemu: 'iso' invalid format https://bugzilla.redhat.com/524111 (lmr) virt-manager can't hold pointer grab on VMs (VNC mode) https://bugzilla.redhat.com/524117 (lmr) virt-manager: Error installing guests - can't mount the / filesystem after the install process is finished https://bugzilla.redhat.com/524118 (P Rauser) NFS4 connection to virtual guest NFS4 server fails over bridged interface https://bugzilla.redhat.com/524205 (vbenes) virt-manager cloning operations should be cancelable Merging Lists ============= Jon Stanley raised the issue again of whether the fedora-virt and fedora-xen mailing lists should be merged: http://www.redhat.com/archives/fedora-virt/2009-October/thread.html#00000 It looks like we'll do this when the lists switch over to the lists.fedoraproject.org server. virt-dostuff ============ Rich Jones announce another couple of tools in the ever growing libguestfs toolsuite, virt-rescue and virt-edit: http://www.redhat.com/archives/fedora-virt/2009-September/msg00099.html Bugs ==== DOOM-O-METER: 198 bugs open now, 214 open three weeks ago. Progress! The Fedora 12 blocker list is: https://bugzilla.redhat.com/showdependencytree.cgi?id=F12VirtBlocker&hide_resolved=1 and the Fedora 12 target list is: https://bugzilla.redhat.com/showdependencytree.cgi?id=F12VirtTarget&hide_resolved=1 All these queries can be found on: https://fedoraproject.org/wiki/Virtualization_bugs Ongoing Bugs ============ == kernel == https://bugzilla.redhat.com/526869 Add virtio_blk support cache flush (VIRTIO_BLK_F_FLUSH) Christoph points out that we should make sure F-12 virtio_blk supports cache flushing for future host versions. https://bugzilla.redhat.com/525890 Backport virtio patches for optimised virtio-net Some debate about whether backporting some changes from 2.6.32 virtio_net is worth the risk. https://bugzilla.redhat.com/524508 kvm regression between 2.6.29 and 2.6.30 causes guest to become unresponsive Reporter has confirmed that his KVM hangs are a regression between 2.6.29 and 2.6.30. https://bugzilla.redhat.com/523900 Garbage in KVM guest console when guest runs kernel 2.6.30.5-43.fc11 A video console under KVM regression between 2.6.29 and 2.6.30. https://bugzilla.redhat.com/524808 kernel-2.6.31-33.fc12.x86_64 fail to boot with VT-d enabled and intel_iommu=on https://bugzilla.redhat.com/490477 [Intel IOMMU] Using isochronous DMAR unit on ICH10 board causes _other_ DMAR unit to stop working. David Woodhouse continues fighting away at VT-d bugs. == misc == https://bugzilla.redhat.com/501720 upstart does not launch a login process on /dev/hvc0 in all cases Since virtio_console is undergoing major changes, we've punted this until Fedora 13. https://bugzilla.redhat.com/492082 Re-phrase anaconda's terrifying "uninitialized drive" warning Looks like anaconda devs just want to close this. https://bugzilla.redhat.com/523296 When Qemu/SDL is fullscreen xrandr dual-monitor configuration breaks Looks like SDL's fullscreen support can somehow screw up xrandr configuration. == qemu == https://bugzilla.redhat.com/524734 KVM guest ext3 errors at shutdown when using virtio and a qcow2 backing file An Ubuntu guest on F12 using a qcow2 backing file and virtio sees ext3 errors. https://bugzilla.redhat.com/526381 qcow2 performance bad under i/o load Quite worrying given tha qcow2 performance is an F-12 feature. However, it looks like this may be just that these are compressed qcow2 which are not expected to have good performance. https://bugzilla.redhat.com/526777 Guest PXE booting doesn't work when using ne2k_pci NIC model https://bugzilla.redhat.com/526776 Add rtl8209 to gpxe-roms-qemu Eduardo points out that we're including the wrong PXE ROM for ne2k_pci. This is also true in F-11. https://bugzilla.redhat.com/523312 KVM guest freezes during DesktopBSD 1.7 installation This issue on a Fedora 11 host sounds like it might be storage related. https://bugzilla.redhat.com/523592 qemu-kvm segfault when attaching USB audio device We have a stack trace for this main loop segfault on F11 with USB passthrough. https://bugzilla.redhat.com/524723 Windows XP KVM gets BSOD with HP all-in-one usb device attached An issue with USB passthrough on Fedora 11. https://bugzilla.redhat.com/503156 qemu VNC :: xterm inside VM shows garbled text A patch destined for 0.10.7 that should have fixed this has, in fact, made things worse. == libvirt == https://bugzilla.redhat.com/524732 org.libvirt.manage policy kit denial Tom Horsley is seeing polkit denials with virsh and virt-maanger. https://bugzilla.redhat.com/517157 libvirt fails to start guest on NFS even when sebool virt_use_nfs is on F12 has a fix to handle EOPNOTSUPP from setfilecon() when running on NFS. Need this in F11 too. https://bugzilla.redhat.com/516430 libvirt cannot re-label a disk image under an NTFS partition Looks like a similar issue to the NFS one, except we don't have a virt_use_ntfs sebool. https://bugzilla.redhat.com/524770 'virsh vol-path' command doesn't support "name" parameter https://bugzilla.redhat.com/524776 Describe vol-key vol-path, vol-name commands wrongly in help doc Some of the storage volume commands seem to be quite messed up ever since they were introduced. https://bugzilla.redhat.com/525073 cannot delete storage pool with virsh pool-delete Another issue with virsh's storage commands. https://bugzilla.redhat.com/526769 libvirt logrotate settings should roll over weekly instead of daily libvirt's logrotate settings are proving to be quite annoying, so we're switching it to only roll-over weekly. == virt-manager == https://bugzilla.redhat.com/522823 virt-manager doc screenshots are outdated The Fedora 12 version of virt-manager has a radically different UI, but the help documentation hasn't been updated yet. https://bugzilla.redhat.com/522720 virt-manager cloning a guest with a 3Gib image creates a 23Gib image virt-manager appears to get confused about image sizes when cloning a guest. https://bugzilla.redhat.com/524297 F11 virt-manager doesn't allow KVM memory ballooning Looks like this is fixed in F12, so it's purely a case of whether we should make the effort to fix in F11 too. https://bugzilla.redhat.com/526077 virt-manager toolbar buttons should have tooltips According to the GNOME HIG, all toolbar buttons should have tooltips, especially ones without a label. https://bugzilla.redhat.com/526076 some virt-manager toolbar icons are blurry Eagle eyed Michael Monreal spotted that some of virt-manager's new icons seem to be blurry due to inappropriate scaling. He even included a nice screenshot demonstrating the problem. https://bugzilla.redhat.com/526111 virt-manager add storage volume allows an empty name field Minor problem with virt-manager not checking the volume name text entry for data. https://bugzilla.redhat.com/526488 virt-manager's 'Details' dialog has a poorly placed 'Remove' button https://bugzilla.redhat.com/526491 virt-manager has no confirmation dialog when deleting a virtual disk Matt Booth accidentally deleted a few guest images because of a combination of these bugs. https://bugzilla.redhat.com/526648 virt-maanger's "New VM" window disappears when connection expander is closed Yet another minor virt-manager UI quirklet. == virt-clone == https://bugzilla.redhat.com/524269 virt-clone should demove old udev rules when changing MAC address Could be implemented using libguestfs. == xen == https://bugzilla.redhat.com/523914 Mouse does not move in PV Xen guest under RHEL-5.4 A rather interesting, complex, twisty bug report detailing why xen and evdev have conspired to break the mouse in Fedora 12 Xen guests. https://bugzilla.redhat.com/523489 32-bit rawhide xen kernel spins 100% cpu booting on CentOS 5.2 Dom0 Xen DomU boot failure. Resolved Bugs ============= == misc == https://bugzilla.redhat.com/507397 Directory permissions on volume group directory too restrictive This blocker issue with lvm2 was first fixed by a large change to using udev, but that broken anaconda so it was reverted, and then it was fixed by a much more minor change. == kernel == https://bugzilla.redhat.com/520119 virtio_net page allocation failure Rusty came up with a fix for this in 2.6.31, so we need to backport it to 2.6.30 in Fedora 11. https://bugzilla.redhat.com/524229 Local migration of kvm guest fails in Fedora12 Alpha A PVMMU bug which has been fixed in 2.6.31. https://bugzilla.redhat.com/509383 rotational mode is much faster for virtio-blk disks, but uses non-rotational mode by default Justin has applied Christoph's fix for this in rawhide and F-11. https://bugzilla.redhat.com/512358 Unable to boot using qemu-kvm and gPXE from virt-preview repository This is now fixed in Fedora 11. == qemu == https://bugzilla.redhat.com/524022 qemu's pci_add monitor command should not exit() if the NIC model is not valid Markus fixed this upstream and the patches have now been pulled into F12. https://bugzilla.redhat.com/524695 qemu should print strerror info for disk/kernel/initrd access errors Cole points out that since we're getting a lot of problems with qemu not being able to access files since we made it run as an unprivileged user, we should really fix qemu's error messages. Justin has sent a patch upstream and included it for F12. https://bugzilla.redhat.com/522683 svirt blocks USB passthrough, even with virt_use_usb enabled - /sys/bus/usb/devices Looks like svirt is breaking USB passthrough by blocking access to /sys/bus/usb/devices. Dan Walsh fixed this in rawhide. https://bugzilla.redhat.com/522994 KVM Fedora 11 guest networking fails with latest (2.6.30.5-43) kernel. F-11 qemu-kvm was missing a crucial piece of GSO support which caused networking to break with 2.6.30 guests. This was only a problem for people using qemu-kvm directly from the command line because libvirt wasn't enabling GSO (#526472). Fixed now in updates-testing. https://bugzilla.redhat.com/509532 Move /usr/bin/qemu-kvm into the qemu-kvm package danpb points out that the qemu-kvm package should go away eventually, so we should keep it as an empty meta-package for comps. https://bugzilla.redhat.com/523677 add support for s390x Dan Hor?k proposed that Ulrich Hecht's patch to add support for s390 to qemu be included in Fedora, but it was decided to just wait for the patch to get upstream. == libvirt == https://bugzilla.redhat.com/524499 libvirt should not enable qemu's audio backend, even with selinux disabled Now that we run qemu as an unprivileged user, we should never enable qemu's sdl/pulse audio backend. A patch to do that from upstream is now in F-12. https://bugzilla.redhat.com/497131 svirt fails to relabel qcow2 backing files This has been fixed by using libvirt's existing image format probing code to determine the backing file and then re-labelling it. Patch has been backported to F-12 now. https://bugzilla.redhat.com/526472 libvirt fails to enable IFF_VNET_HDR on tap fds for qemu-kvm In F-11 libvirt was failing to recognise that qemu had GSO support; fixed by backporting patches to re-work how libvirt probes qemu versions. https://bugzilla.redhat.com/523766 AVCs from libvirtd startup A bunch of SELinux AVCs during libvirtd startup that look like they might be fixed by latest upstream netcf. Fixed by Dan Walsh in latest rawhide. https://bugzilla.redhat.com/517617 libvirt/netcf loads modprobe.conf and others - AVC messages (preventing libvirtd (virtd_t) "getattr" modules_conf_t) More netcf related AVCs which have been fixed. https://bugzilla.redhat.com/520864 libvirt is using untrusted 'info vcpus' PID data for already running VM after libvirtd restart https://bugzilla.redhat.com/465532 RFE: libvirt should support KVM huge page backed memory https://bugzilla.redhat.com/517619 libvirtd should chown dirs when qemu configured to run as root/root All fixed in rawhide by libvirt-0.7.1. == python-virtinst == https://bugzilla.redhat.com/523960 virtinst saves images where qemu can't access them by default Cole has fixed this by making virtinst added qemu to the ACLs for the directories qemu will need access to. https://bugzilla.redhat.com/524109 virt-manager: Fails to install guest using ISO file - internal error unable to start guest: qemu: 'iso' invalid format The fix for bug #517151 introduces another problem; Cole has now fixed that one in rawhide. https://bugzilla.redhat.com/499654 failure to add more than 16 virtio-blk devices in virt-manager https://bugzilla.redhat.com/505317 virtinst: make SLES11 guests use virtio by default https://bugzilla.redhat.com/506319 virtinst errors finding default bridge: upsets virt-manager Various F11 virtinst bugs fixed by python-virtinst-0.400.3-9.fc11. == virt-manager == https://bugzilla.redhat.com/517379 virt-manager should warn if guest images will are not readable by qemu When attaching any storage to a guest we now attempt to verify that the 'qemu' user has search access. If not, we warn the user and offer to fix, using ACLs. The user can opt out, and optionally 'Never ask about this path again" if we have any false positives. https://bugzilla.redhat.com/517664 virt-manager ignores "Host does not support any virtualization options" error Cole has made this error condition a bit more obvious to the user. https://bugzilla.redhat.com/522823 virt-manager doc screenshots are outdated The help docs are so woefully out of date, Cole has just disabled them for now. https://bugzilla.redhat.com/502777 virt-manager does not recalculate free disk space on new VM creation While the 'New VM' dialog is up, virt-manager now polls every 3 seconds to check the available disk space on the host. From ddutile at redhat.com Fri Oct 2 18:06:59 2009 From: ddutile at redhat.com (Don Dutile) Date: Fri, 02 Oct 2009 14:06:59 -0400 Subject: [fedora-virt] Fedora virt status In-Reply-To: <1254505706.4346.18.camel@blaa> References: <1254505706.4346.18.camel@blaa> Message-ID: <4AC64143.3010408@redhat.com> Mark McLoughlin wrote: > Fedora 12 > ========= > > The Fedora 12 release is drawing closer and closer. Beta is on its way > out the door and release candidate composes will begin in less than > four weeks time. > > Helping Out > =========== > > Here are three ways you could help out with getting F12 into great > shape: > > 1) Test, test, test and file bugs. The more stuff you break, the > more will get fixed. > > Maybe you could even write some test cases; for examples, just > take a look at: > > https://fedoraproject.org/wiki/Test_Day:2009-09-17_Virtualization > https://fedoraproject.org/wiki/Test_Day:2009-05-07_Virtualization > > 2) Follow the action in bugzilla. See this wiki page for how to sign > up to receive all bugzilla emails relating to Fedora > virtualization: > > https://fedoraproject.org/wiki/Virtualization_bugs > > 3) Take a look at some existing bugs and see if you can help out > getting them closed. A good place to start is the F12 target > list: > > https://bugzilla.redhat.com/showdependencytree.cgi?id=F12VirtTarget&hide_resolved=1 > > But be warned, you may quickly find yourself becoming a package > co-maintainer! :-) > You really know how to sell it to the troops! ;-) fore-warned is fore-armed.... move out soldier! ... Don From sven at lank.es Sat Oct 3 13:44:18 2009 From: sven at lank.es (Sven Lankes) Date: Sat, 3 Oct 2009 15:44:18 +0200 Subject: [fedora-virt] virt-manager and manually created bridges Message-ID: <20091003134417.GC6494@killefiz> My home-server is running rawhide. To get some virtual machines working on it I did: * disable Networkmanager * setup a br0-Device that includes eth0 * disable the libvirt default network virbr0 using virsh net-autostart default --disable When I now try to install a new vm using virt-manager this fails: I'm asked (under advanced options) to choose a network - the only network available ist the (deactivated) virbr0 network. When I then try to proceed anyway I am told that the virbr0 is inactive and would I like to start it. Choosing no doesn't allow me to finish the installation wizard. When I edit an xml definition manually to add the br0-Bridge as network device things work fine. Shouldn't this be possible from within virt-manager? -- sven === jabber/xmpp: sven at lankes.net From tom.horsley at att.net Sat Oct 3 15:16:53 2009 From: tom.horsley at att.net (Tom Horsley) Date: Sat, 3 Oct 2009 11:16:53 -0400 Subject: [fedora-virt] virt-manager and manually created bridges In-Reply-To: <20091003134417.GC6494@killefiz> References: <20091003134417.GC6494@killefiz> Message-ID: <20091003111653.524f7a71@zooty> On Sat, 3 Oct 2009 15:44:18 +0200 Sven Lankes wrote: > * disable Networkmanager Do you also enable "network"? > * setup a br0-Device that includes eth0 > * disable the libvirt default network virbr0 using virsh net-autostart > default --disable I do almost the same thing, but I use virsh to undefine the default network completely, when I install a vm, it just automatically offers br0 as the only choice for network. Here's my ifcfg-br0 and ifcfg-eth0 files (I use a static IP): DEVICE=br0 TYPE=Bridge BOOTPROTO=static GATEWAY=192.168.1.1 IPADDR=192.168.1.106 IPV6INIT=no NETMASK=255.255.255.0 ONBOOT=yes PEERDNS=no USERCTL=no NM_CONTROLLED=no DEVICE=eth0 HWADDR=00:1b:21:3a:00:f3 ONBOOT=yes TYPE=Ethernet IPV6INIT=no USERCTL=no BRIDGE=br0 NM_CONTROLLED=no PEERDNS=no From tom.horsley at att.net Sat Oct 3 19:38:23 2009 From: tom.horsley at att.net (Tom Horsley) Date: Sat, 03 Oct 2009 19:38:23 +0000 Subject: [fedora-virt] user qemu has no access to usb devices Message-ID: <100320091938.13024.4AC7A82F000BA317000032E022216125569B0A02D29B9B0EBF970A049C9D0108D203019B@att.net> I don't know if something is busted in my rawhide setup, or if this is just a real problem with the new scheme of running as user qemu, but user qemu (on my system at least), has no access to physical usb devices. I had to change the qemu.conf file to tell it to run as root before it could attach a usb device to my windows XP virtual machine. See https://bugzilla.redhat.com/show_bug.cgi?id=524723#c5 From rampal.anuj at gmail.com Mon Oct 5 09:27:14 2009 From: rampal.anuj at gmail.com (anuj rampal) Date: Mon, 5 Oct 2009 14:57:14 +0530 Subject: [fedora-virt] virt-viewer windows port In-Reply-To: <20091001084646.GB32442@amd.home.annexia.org> References: <20090923145706.GA31875@amd.home.annexia.org> <20090924081844.GA32184@amd.home.annexia.org> <20090924132443.GB32184@amd.home.annexia.org> <20090925120254.GC32184@amd.home.annexia.org> <20090930162125.GD32184@amd.home.annexia.org> <20091001084646.GB32442@amd.home.annexia.org> Message-ID: Hi, But what were the messages in the configure step? As i said that virt-viewer takes the certificates from "Z:\usr\i686-pc-mingw32\sys-root\mingw\etc\pki..." path in windows. This is what i tried: 1. when I do not keep the certificates on this path. This is the result that I get. URI: virt-viewer.exe -c qemu://FC11-KVM/session 1 This is the result: 14:53:35.044: debug : virInitialize:294 : register drivers 14:53:35.050: debug : virRegisterDriver:735 : registering Test as driver 0 14:53:35.051: debug : virRegisterNetworkDriver:604 : registering Test as network driver 0 14:53:35.051: debug : virRegisterInterfaceDriver:635 : registering Test as inter face driver 0 14:53:35.051: debug : virRegisterStorageDriver:666 : registering Test as storage driver 0 14:53:35.052: debug : virRegisterDeviceMonitor:697 : registering Test as device driver 0 14:53:35.052: debug : virRegisterDriver:735 : registering ESX as driver 1 14:53:35.052: debug : virRegisterDriver:735 : registering remote as driver 2 14:53:35.052: debug : virRegisterNetworkDriver:604 : registering remote as netwo rk driver 1 14:53:35.053: debug : virRegisterInterfaceDriver:635 : registering remote as int erface driver 1 14:53:35.053: debug : virRegisterStorageDriver:666 : registering remote as stora ge driver 1 14:53:35.053: debug : virRegisterDeviceMonitor:697 : registering remote as devic e driver 1 14:53:35.070: debug : virConnectOpenAuth:1214 : name=qemu://FC11-KVM/session, au th=0022FD88, flags=1 14:53:35.072: debug : do_open:1001 : name "qemu://FC11-KVM/session" to URI compo nents: scheme qemu opaque (null) authority (null) server FC11-KVM user (null) port 0 path /session 14:53:35.077: debug : do_open:1011 : trying driver 0 (Test) ... 14:53:35.078: debug : do_open:1017 : driver 0 Test returned DECLINED 14:53:35.083: debug : do_open:1011 : trying driver 1 (ESX) ... 14:53:35.084: debug : do_open:1017 : driver 1 ESX returned DECLINED 14:53:35.086: debug : do_open:1011 : trying driver 2 (remote) ... 14:53:35.087: debug : doRemoteOpen:533 : proceeding with name = qemu:///session 14:53:35.104: debug : do_open:1017 : driver 2 remote returned ERROR 14:53:35.105: debug : virUnrefConnect:232 : unref connection 02ACA3A0 1 14:53:35.107: debug : virReleaseConnect:191 : release connection 02ACA3A0 unable to connect to libvirt qemu://FC11-KVM/session 2. When i again palce the certificates at this path. This is the result that i get. URI: virt-viewer.exe -c qemu://FC11-KVM/session 1 This is the result: 14:56:02.938: debug : virInitialize:294 : register drivers 14:56:02.944: debug : virRegisterDriver:735 : registering Test as driver 0 14:56:02.944: debug : virRegisterNetworkDriver:604 : registering Test as network driver 0 14:56:02.944: debug : virRegisterInterfaceDriver:635 : registering Test as inter face driver 0 14:56:02.945: debug : virRegisterStorageDriver:666 : registering Test as storage driver 0 14:56:02.949: debug : virRegisterDeviceMonitor:697 : registering Test as device driver 0 14:56:02.950: debug : virRegisterDriver:735 : registering ESX as driver 1 14:56:02.950: debug : virRegisterDriver:735 : registering remote as driver 2 14:56:02.950: debug : virRegisterNetworkDriver:604 : registering remote as netwo rk driver 1 14:56:02.951: debug : virRegisterInterfaceDriver:635 : registering remote as int erface driver 1 14:56:02.951: debug : virRegisterStorageDriver:666 : registering remote as stora ge driver 1 14:56:02.951: debug : virRegisterDeviceMonitor:697 : registering remote as devic e driver 1 14:56:02.960: debug : virConnectOpenAuth:1214 : name=qemu://FC11-KVM/session, au th=0022FD88, flags=1 14:56:02.962: debug : do_open:1001 : name "qemu://FC11-KVM/session" to URI compo nents: scheme qemu opaque (null) authority (null) server FC11-KVM user (null) port 0 path /session 14:56:02.966: debug : do_open:1011 : trying driver 0 (Test) ... 14:56:02.970: debug : do_open:1017 : driver 0 Test returned DECLINED 14:56:02.972: debug : do_open:1011 : trying driver 1 (ESX) ... 14:56:02.973: debug : do_open:1017 : driver 1 ESX returned DECLINED 14:56:02.974: debug : do_open:1011 : trying driver 2 (remote) ... 14:56:02.976: debug : doRemoteOpen:533 : proceeding with name = qemu:///session 14:56:02.993: debug : initialise_gnutls:1099 : loading CA file /usr/i686-pc-ming w32/sys-root/mingw/etc/pki/CA/cacert.pem 14:56:02.996: debug : initialise_gnutls:1112 : loading client cert and key from files /usr/i686-pc-mingw32/sys-root/mingw/etc/pki/libvirt/clientcert.pem and /us r/i686-pc-mingw32/sys-root/mingw/etc/pki/libvirt/private/clientkey.pem 14:56:03.114: debug : do_open:1017 : driver 2 remote returned ERROR 14:56:03.115: debug : virUnrefConnect:232 : unref connection 025EA3A0 1 14:56:03.117: debug : virReleaseConnect:191 : release connection 025EA3A0 unable to connect to libvirt qemu://FC11-KVM/session Regards Anuj On Thu, Oct 1, 2009 at 2:16 PM, Richard W.M. Jones wrote: > On Thu, Oct 01, 2009 at 02:00:49PM +0530, anuj rampal wrote: > > > i tried "./configure --sysconfdir=C:\pki\" > > > but this didnot work... > > > > When you say this "did not work", how did it fail? The clue is > > usually in the error message. > > > > when i did --sysconfdir="C:\pki\" I expected that i have keep my > > certificates under this directory(correct me if i'm wrong). > > But it was still taking the certificates from > > "Z:\usr\i686-pc-mingw\sys-confic\mingw\etc\pki" (in Windows). > > But what were the messages in the configure step? > > Rich. > > -- > Richard Jones, Emerging Technologies, Red Hat > http://et.redhat.com/~rjones > virt-df lists disk usage of guests without needing to install any > software inside the virtual machine. Supports Linux and Windows. > http://et.redhat.com/~rjones/virt-df/ > -------------- next part -------------- An HTML attachment was scrubbed... URL: From rjones at redhat.com Mon Oct 5 17:23:13 2009 From: rjones at redhat.com (Richard W.M. Jones) Date: Mon, 5 Oct 2009 18:23:13 +0100 Subject: [fedora-virt] [ANNOUNCE] virt-top 1.0.4 Message-ID: <20091005172313.GA9574@amd.home.annexia.org> virt-top is the sane replacement for xentop. Virt-top looks and acts like the familiar top(1) command, displays virtual machines, and uses libvirt so it works with just about every virtualization system out there. It also has cool features for sysadmins, like you can use it to log stats into a database or spreadsheet. This version fixes a few long-standing bugs, particularly in the handling of languages other than English. We've also tidied up the source and moved to using git for source control. Home page: http://et.redhat.com/~rjones/virt-top/ Source: http://et.redhat.com/~rjones/virt-top/files/ Repo: http://git.et.redhat.com/?p=virt-top.git;a=summary My blog: http://rwmj.wordpress.com/tag/virt-top/ There's already a build in Fedora 13 which will probably work in earlier versions of Fedora: http://koji.fedoraproject.org/koji/packageinfo?packageID=6211 Rich. -- Richard Jones, Emerging Technologies, Red Hat http://et.redhat.com/~rjones virt-df lists disk usage of guests without needing to install any software inside the virtual machine. Supports Linux and Windows. http://et.redhat.com/~rjones/virt-df/ From wildfire at progsoc.org Mon Oct 5 23:31:35 2009 From: wildfire at progsoc.org (Anand Kumria) Date: Tue, 6 Oct 2009 00:31:35 +0100 Subject: [fedora-virt] virt-viewer windows port In-Reply-To: References: <20090923145706.GA31875@amd.home.annexia.org> <20090924132443.GB32184@amd.home.annexia.org> <20090925120254.GC32184@amd.home.annexia.org> <20090930162125.GD32184@amd.home.annexia.org> <20091001084646.GB32442@amd.home.annexia.org> Message-ID: <971f65790910051631i4a03792fua6379e8fb278e7f9@mail.gmail.com> Hi, What's with the colours? Why are you making it harder to understand what was said by who? On Mon, Oct 5, 2009 at 10:27 AM, anuj rampal wrote: > Hi, > > It would also be really useful if you kept the attribution lines, so we could all see who had said what, when. What Richard is asking you for is the output of the ./configure command. What he hasn't asked, yet, is how porting libvirt and the virt-* utilis helps you. What is the actual end goal here? A "just for fun" experiment, in which case you are probably better served by answers to your questions which lead you to do the deeper debugging. Or a school project, in which case the same applies. Thanks, Anand -------------- next part -------------- An HTML attachment was scrubbed... URL: From rampal.anuj at gmail.com Tue Oct 6 06:32:58 2009 From: rampal.anuj at gmail.com (anuj rampal) Date: Tue, 6 Oct 2009 12:02:58 +0530 Subject: [fedora-virt] virt-viewer windows port In-Reply-To: <971f65790910051631i4a03792fua6379e8fb278e7f9@mail.gmail.com> References: <20090923145706.GA31875@amd.home.annexia.org> <20090924132443.GB32184@amd.home.annexia.org> <20090925120254.GC32184@amd.home.annexia.org> <20090930162125.GD32184@amd.home.annexia.org> <20091001084646.GB32442@amd.home.annexia.org> <971f65790910051631i4a03792fua6379e8fb278e7f9@mail.gmail.com> Message-ID: On Tue, Oct 6, 2009 at 5:01 AM, Anand Kumria wrote: > Hi, > > What's with the colours? Why are you making it harder to understand what > was said by who? > It would also be really useful if you kept the attribution lines, so we > could all see who had said what, when. > > What Richard is asking you for is the output of the ./configure command. > These are the packages that are required: as given in the mingw32.spec file mingw32-filesystem mingw32-gtk2 mingw32-libvirt mingw32-libxml2 mingw32-libglade2 mingw32-gtk-vnc pkgconfig These are all there.. [root at FC11-KVM ~]# rpm -qa | grep mingw32-filesystem mingw32-filesystem-50-3.fc11.1.noarch [root at FC11-KVM ~]# rpm -qa | grep mingw32-gtk2 mingw32-gtk2-2.16.6-1.fc11.noarch mingw32-gtk2-static-2.16.6-1.fc11.noarch [root at FC11-KVM ~]# rpm -qa | grep mingw32-libvirt mingw32-libvirt-0.6.1-1.fc11.noarch [root at FC11-KVM ~]# rpm -qa | grep mingw32-libxml2 mingw32-libxml2-2.7.4-1.fc11.noarch mingw32-libxml2-static-2.7.4-1.fc11.noarch [root at FC11-KVM ~]# rpm -qa | grep mingw32-libglade2 mingw32-libglade2-2.6.4-2.fc11.noarch mingw32-libglade2-static-2.6.4-2.fc11.noarch [root at FC11-KVM ~]# rpm -qa | grep mingw32-gtk-vnc mingw32-gtk-vnc-0.3.8-5.fc11.noarch [root at FC11-KVM ~]# rpm -qa | grep pkgconfig pkgconfig-0.23-8.fc11.i586 ----------------------- this is how i configured the virt-viewer: PKG_CONFIG_PATH="/usr/i686-pc-mingw32/sys-root/mingw/lib/pkgconfig" CC="i686-pc-mingw32-gcc" ./configure --build=i386-pc-linux --host=i686-pc-mingw32 --prefix="/usr/i686-pc-mingw32/sys-root/mingw" --sysconfdir=C:\\pki and here is the result: checking for a BSD-compatible install... /usr/bin/install -c checking whether build environment is sane... yes checking for i686-pc-mingw32-strip... i686-pc-mingw32-strip checking for a thread-safe mkdir -p... /bin/mkdir -p checking for gawk... gawk checking whether make sets $(MAKE)... yes checking build system type... i386-pc-linux-gnu checking host system type... i686-pc-mingw32 checking for i686-pc-mingw32-gcc... i686-pc-mingw32-gcc checking for C compiler default output file name... a.exe checking whether the C compiler works... yes checking whether we are cross compiling... yes checking for suffix of executables... .exe checking for suffix of object files... o checking whether we are using the GNU C compiler... yes checking whether i686-pc-mingw32-gcc accepts -g... yes checking for i686-pc-mingw32-gcc option to accept ISO C89... none needed checking for style of include used by make... GNU checking dependency style of i686-pc-mingw32-gcc... gcc3 checking whether i686-pc-mingw32-gcc and cc understand -c and -o together... yes checking for a sed that does not truncate output... /bin/sed checking for grep that handles long lines and -e... /bin/grep checking for egrep... /bin/grep -E checking for fgrep... /bin/grep -F checking for ld used by i686-pc-mingw32-gcc... /usr/i686-pc-mingw32/bin/ld checking if the linker (/usr/i686-pc-mingw32/bin/ld) is GNU ld... yes checking for BSD- or MS-compatible name lister (nm)... /usr/bin/i686-pc-mingw32-nm -B checking the name lister (/usr/bin/i686-pc-mingw32-nm -B) interface... BSD nm checking whether ln -s works... yes checking the maximum length of command line arguments... 1966080 checking whether the shell understands some XSI constructs... yes checking whether the shell understands "+="... yes checking for /usr/i686-pc-mingw32/bin/ld option to reload object files... -r checking for i686-pc-mingw32-objdump... i686-pc-mingw32-objdump checking how to recognize dependent libraries... file_magic ^x86 archive import|^x86 DLL checking for i686-pc-mingw32-ar... i686-pc-mingw32-ar checking for i686-pc-mingw32-strip... (cached) i686-pc-mingw32-strip checking for i686-pc-mingw32-ranlib... i686-pc-mingw32-ranlib checking command to parse /usr/bin/i686-pc-mingw32-nm -B output from i686-pc-mingw32-gcc object... ok checking how to run the C preprocessor... i686-pc-mingw32-gcc -E checking for ANSI C header files... yes checking for sys/types.h... yes checking for sys/stat.h... yes checking for stdlib.h... yes checking for string.h... yes checking for memory.h... yes checking for strings.h... yes checking for inttypes.h... yes checking for stdint.h... yes checking for unistd.h... yes checking for dlfcn.h... yes checking for objdir... .libs checking if i686-pc-mingw32-gcc supports -fno-rtti -fno-exceptions... no checking for i686-pc-mingw32-gcc option to produce PIC... -DDLL_EXPORT -DPIC checking if i686-pc-mingw32-gcc PIC flag -DDLL_EXPORT -DPIC works... yes checking if i686-pc-mingw32-gcc static flag -static works... yes checking if i686-pc-mingw32-gcc supports -c -o file.o... yes checking if i686-pc-mingw32-gcc supports -c -o file.o... (cached) yes checking whether the i686-pc-mingw32-gcc linker (/usr/i686-pc-mingw32/bin/ld) supports shared libraries... yes checking whether -lc should be explicitly linked in... yes checking dynamic linker characteristics... Win32 ld.exe checking how to hardcode library paths into programs... immediate checking whether stripping libraries is possible... yes checking if libtool supports shared libraries... yes checking whether to build shared libraries... yes checking whether to build static libraries... yes checking whether gcc understands -Wp,-D_FORTIFY_SOURCE=2... yes checking whether gcc understands -fexceptions... yes checking whether gcc understands -fstack-protector... yes checking whether gcc understands --param=ssp-buffer-size=4... yes checking whether gcc understands -fasynchronous-unwind-tables... yes checking whether gcc understands -Wall... yes checking whether gcc understands -Wmissing-prototypes... yes checking whether gcc understands -std=c99... yes checking whether gcc understands -Wnested-externs... yes checking whether gcc understands -Wpointer-arith... yes checking whether gcc understands -Wextra... yes checking whether gcc understands -Wshadow... yes checking whether gcc understands -Wcast-align... yes checking whether gcc understands -Wwrite-strings... yes checking whether gcc understands -Waggregate-return... yes checking whether gcc understands -Winline... yes checking whether gcc understands -Wredundant-decls... yes checking whether gcc understands -Wno-sign-compare... yes checking what language compliance flags to pass to the C compiler... checking for i686-pc-mingw32-pkg-config... no checking for pkg-config... /usr/bin/pkg-config configure: WARNING: using cross tools not prefixed with host triplet checking pkg-config is at least version 0.9.0... yes checking for LIBXML2... yes checking for LIBVIRT... yes checking for GTK2... yes checking for LIBGLADE2... yes checking for GTKVNC... yes checking sys/socket.h usability... no checking sys/socket.h presence... no checking for sys/socket.h... no checking sys/un.h usability... no checking sys/un.h presence... no checking for sys/un.h... no checking windows.h usability... yes checking windows.h presence... yes checking for windows.h... yes checking for fork... no checking for socketpair... no configure: creating ./config.status config.status: creating Makefile config.status: creating src/Makefile config.status: creating man/Makefile config.status: creating plugin/Makefile config.status: creating virt-viewer.spec config.status: creating mingw32-virt-viewer.spec config.status: creating config.h config.status: executing depfiles commands config.status: executing libtool commands > > What he hasn't asked, yet, is how porting libvirt and the virt-* utilis > helps you. > > What is the actual end goal here? A "just for fun" experiment, in which > case you are probably better served by answers to your questions which lead > you to do the deeper debugging. Or a school project, in which case the same > applies. > > Thanks, > Anand > I had to call the libvirt functions from my windows machine and this could only be done by porting libvirt to windows(as a client).. i have done that sucessfully and now I can call all the functions from my windows machine... Now what happens is that after I create a new Guest OS using my code(from windows), I have to logon into my server machine and then continue installing the Guest. This is where i thought of porting virt-viewer to windows so that i can do it from the same remote windows machine. The only problem that im facing right now is with the virt-viewer. If this issues is resolved, we are thinking of intergate libvirt with one of our product. Regards Anuj -------------- next part -------------- An HTML attachment was scrubbed... URL: From rampal.anuj at gmail.com Tue Oct 6 11:02:44 2009 From: rampal.anuj at gmail.com (anuj rampal) Date: Tue, 6 Oct 2009 16:32:44 +0530 Subject: [fedora-virt] virt-viewer windows port In-Reply-To: References: <20090923145706.GA31875@amd.home.annexia.org> <20090925120254.GC32184@amd.home.annexia.org> <20090930162125.GD32184@amd.home.annexia.org> <20091001084646.GB32442@amd.home.annexia.org> <971f65790910051631i4a03792fua6379e8fb278e7f9@mail.gmail.com> Message-ID: Hi, I think there is some problem the windows port of virt-viewer or i guess i have done something wrong. I have made the Guest OS display on my Windows Machine but its not through virt-viewer it just by using vnc. As i was able to connect from a remote linux machine using virt-viewer. I figured out that it connects to the libvirtd first and the connects to the display port of the Guest OS. Now when i was trying to connect from my windows machine, what was happening was it was connecting to libvirtd and the hungs up and doesnot connect to the display port of the Guest OS. So i tried to directly connect to the display port of that Guest OS from my Windows machine and i was able to see my Guest OS. I dont know why virt-viewer is not working properly because internally it also uses GTK-VNC only...???? Thanks & Regards Anuj On Tue, Oct 6, 2009 at 12:02 PM, anuj rampal wrote: > On Tue, Oct 6, 2009 at 5:01 AM, Anand Kumria wrote: > >> Hi, >> >> What's with the colours? Why are you making it harder to understand what >> was said by who? > > >> It would also be really useful if you kept the attribution lines, so we >> could all see who had said what, when. >> >> What Richard is asking you for is the output of the ./configure command. >> > > > > These are the packages that are required: as given in the mingw32.spec > file > mingw32-filesystem > mingw32-gtk2 > mingw32-libvirt > mingw32-libxml2 > mingw32-libglade2 > mingw32-gtk-vnc > pkgconfig > > These are all there.. > > [root at FC11-KVM ~]# rpm -qa | grep mingw32-filesystem > mingw32-filesystem-50-3.fc11.1.noarch > [root at FC11-KVM ~]# rpm -qa | grep mingw32-gtk2 > mingw32-gtk2-2.16.6-1.fc11.noarch > mingw32-gtk2-static-2.16.6-1.fc11.noarch > [root at FC11-KVM ~]# rpm -qa | grep mingw32-libvirt > mingw32-libvirt-0.6.1-1.fc11.noarch > [root at FC11-KVM ~]# rpm -qa | grep mingw32-libxml2 > mingw32-libxml2-2.7.4-1.fc11.noarch > mingw32-libxml2-static-2.7.4-1.fc11.noarch > [root at FC11-KVM ~]# rpm -qa | grep mingw32-libglade2 > mingw32-libglade2-2.6.4-2.fc11.noarch > mingw32-libglade2-static-2.6.4-2.fc11.noarch > [root at FC11-KVM ~]# rpm -qa | grep mingw32-gtk-vnc > mingw32-gtk-vnc-0.3.8-5.fc11.noarch > [root at FC11-KVM ~]# rpm -qa | grep pkgconfig > pkgconfig-0.23-8.fc11.i586 > > > ----------------------- > > this is how i configured the virt-viewer: > > PKG_CONFIG_PATH="/usr/i686-pc-mingw32/sys-root/mingw/lib/pkgconfig" > CC="i686-pc-mingw32-gcc" ./configure --build=i386-pc-linux > --host=i686-pc-mingw32 --prefix="/usr/i686-pc-mingw32/sys-root/mingw" > --sysconfdir=C:\\pki > > > and here is the result: > > > checking for a BSD-compatible install... /usr/bin/install -c > checking whether build environment is sane... yes > checking for i686-pc-mingw32-strip... i686-pc-mingw32-strip > checking for a thread-safe mkdir -p... /bin/mkdir -p > checking for gawk... gawk > checking whether make sets $(MAKE)... yes > checking build system type... i386-pc-linux-gnu > checking host system type... i686-pc-mingw32 > checking for i686-pc-mingw32-gcc... i686-pc-mingw32-gcc > checking for C compiler default output file name... a.exe > checking whether the C compiler works... yes > checking whether we are cross compiling... yes > checking for suffix of executables... .exe > checking for suffix of object files... o > checking whether we are using the GNU C compiler... yes > checking whether i686-pc-mingw32-gcc accepts -g... yes > checking for i686-pc-mingw32-gcc option to accept ISO C89... none needed > checking for style of include used by make... GNU > checking dependency style of i686-pc-mingw32-gcc... gcc3 > checking whether i686-pc-mingw32-gcc and cc understand -c and -o > together... yes > checking for a sed that does not truncate output... /bin/sed > checking for grep that handles long lines and -e... /bin/grep > checking for egrep... /bin/grep -E > checking for fgrep... /bin/grep -F > checking for ld used by i686-pc-mingw32-gcc... /usr/i686-pc-mingw32/bin/ld > checking if the linker (/usr/i686-pc-mingw32/bin/ld) is GNU ld... yes > checking for BSD- or MS-compatible name lister (nm)... > /usr/bin/i686-pc-mingw32-nm -B > checking the name lister (/usr/bin/i686-pc-mingw32-nm -B) interface... BSD > nm > checking whether ln -s works... yes > checking the maximum length of command line arguments... 1966080 > checking whether the shell understands some XSI constructs... yes > checking whether the shell understands "+="... yes > checking for /usr/i686-pc-mingw32/bin/ld option to reload object files... > -r > checking for i686-pc-mingw32-objdump... i686-pc-mingw32-objdump > checking how to recognize dependent libraries... file_magic ^x86 archive > import|^x86 DLL > checking for i686-pc-mingw32-ar... i686-pc-mingw32-ar > checking for i686-pc-mingw32-strip... (cached) i686-pc-mingw32-strip > checking for i686-pc-mingw32-ranlib... i686-pc-mingw32-ranlib > checking command to parse /usr/bin/i686-pc-mingw32-nm -B output from > i686-pc-mingw32-gcc object... ok > checking how to run the C preprocessor... i686-pc-mingw32-gcc -E > checking for ANSI C header files... yes > checking for sys/types.h... yes > checking for sys/stat.h... yes > checking for stdlib.h... yes > checking for string.h... yes > checking for memory.h... yes > checking for strings.h... yes > checking for inttypes.h... yes > checking for stdint.h... yes > checking for unistd.h... yes > checking for dlfcn.h... yes > checking for objdir... .libs > checking if i686-pc-mingw32-gcc supports -fno-rtti -fno-exceptions... no > checking for i686-pc-mingw32-gcc option to produce PIC... -DDLL_EXPORT > -DPIC > checking if i686-pc-mingw32-gcc PIC flag -DDLL_EXPORT -DPIC works... yes > checking if i686-pc-mingw32-gcc static flag -static works... yes > checking if i686-pc-mingw32-gcc supports -c -o file.o... yes > checking if i686-pc-mingw32-gcc supports -c -o file.o... (cached) yes > checking whether the i686-pc-mingw32-gcc linker > (/usr/i686-pc-mingw32/bin/ld) supports shared libraries... yes > checking whether -lc should be explicitly linked in... yes > checking dynamic linker characteristics... Win32 ld.exe > checking how to hardcode library paths into programs... immediate > checking whether stripping libraries is possible... yes > checking if libtool supports shared libraries... yes > checking whether to build shared libraries... yes > checking whether to build static libraries... yes > checking whether gcc understands -Wp,-D_FORTIFY_SOURCE=2... yes > checking whether gcc understands -fexceptions... yes > checking whether gcc understands -fstack-protector... yes > checking whether gcc understands --param=ssp-buffer-size=4... yes > checking whether gcc understands -fasynchronous-unwind-tables... yes > checking whether gcc understands -Wall... yes > checking whether gcc understands -Wmissing-prototypes... yes > checking whether gcc understands -std=c99... yes > checking whether gcc understands -Wnested-externs... yes > checking whether gcc understands -Wpointer-arith... yes > checking whether gcc understands -Wextra... yes > checking whether gcc understands -Wshadow... yes > checking whether gcc understands -Wcast-align... yes > checking whether gcc understands -Wwrite-strings... yes > checking whether gcc understands -Waggregate-return... yes > checking whether gcc understands -Winline... yes > checking whether gcc understands -Wredundant-decls... yes > checking whether gcc understands -Wno-sign-compare... yes > checking what language compliance flags to pass to the C compiler... > checking for i686-pc-mingw32-pkg-config... no > checking for pkg-config... /usr/bin/pkg-config > configure: WARNING: using cross tools not prefixed with host triplet > checking pkg-config is at least version 0.9.0... yes > checking for LIBXML2... yes > checking for LIBVIRT... yes > checking for GTK2... yes > checking for LIBGLADE2... yes > checking for GTKVNC... yes > checking sys/socket.h usability... no > checking sys/socket.h presence... no > checking for sys/socket.h... no > checking sys/un.h usability... no > checking sys/un.h presence... no > checking for sys/un.h... no > checking windows.h usability... yes > checking windows.h presence... yes > checking for windows.h... yes > checking for fork... no > checking for socketpair... no > configure: creating ./config.status > config.status: creating Makefile > config.status: creating src/Makefile > config.status: creating man/Makefile > config.status: creating plugin/Makefile > config.status: creating virt-viewer.spec > config.status: creating mingw32-virt-viewer.spec > config.status: creating config.h > config.status: executing depfiles commands > config.status: executing libtool commands > > > >> >> What he hasn't asked, yet, is how porting libvirt and the virt-* utilis >> helps you. >> >> What is the actual end goal here? A "just for fun" experiment, in which >> case you are probably better served by answers to your questions which lead >> you to do the deeper debugging. Or a school project, in which case the same >> applies. >> >> Thanks, >> Anand >> > > > I had to call the libvirt functions from my windows machine and this could > only be done by porting libvirt to windows(as a client).. i have done that > sucessfully and now I can call all the functions from my windows machine... > > Now what happens is that after I create a new Guest OS using my code(from > windows), I have to logon into my server machine and then continue > installing the Guest. This is where i thought of porting virt-viewer to > windows so that i can do it from the same remote windows machine. > > The only problem that im facing right now is with the virt-viewer. If this > issues is resolved, we are thinking of intergate libvirt with one of our > product. > > Regards > Anuj > -------------- next part -------------- An HTML attachment was scrubbed... URL: From mrichytech at googlemail.com Mon Oct 5 22:45:09 2009 From: mrichytech at googlemail.com (Tim Richardson) Date: Mon, 5 Oct 2009 23:45:09 +0100 Subject: [fedora-virt] Domain doesn't show up Message-ID: I'm trying to install xp virtual machine from an .iso on the disk, I have a virtual machine called localhost with qemu as ID. It has a storage area at /var/lib/libvirt/images, i'm right clicking on that and choosing new then following the wizard but it keeps failing whilst creating the domain. Is anyone familiar with this scenario or have any suggestions? regards Tim Unable to complete install ' internal error Domain microXP didn't show up Traceback (most recent call last): File "/usr/share/virt-manager/virtManager/create.py", line 1501, in do_install dom = guest.start_install(False, meter = meter) File "/usr/lib/python2.6/site-packages/virtinst/Guest.py", line 541, in start_install return self._do_install(consolecb, meter, removeOld, wait) File "/usr/lib/python2.6/site-packages/virtinst/Guest.py", line 633, in _do_install self.domain = self.conn.createLinux(install_xml, 0) File "/usr/lib64/python2.6/site-packages/libvirt.py", line 974, in createLinux if ret is None:raise libvirtError('virDomainCreateLinux() failed', conn=self) libvirtError: internal error Domain microXP didn't show up ' -------------- next part -------------- An HTML attachment was scrubbed... URL: From markmc at redhat.com Tue Oct 6 13:04:35 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Tue, 06 Oct 2009 14:04:35 +0100 Subject: [fedora-virt] virt-preview updates In-Reply-To: <1250703309.17874.46.camel@blaa> References: <1244034684.5001.134.camel@blaa> <1244046483.5001.180.camel@blaa> <1244197459.27876.0.camel@blaa> <1246303956.11688.148.camel@blaa> <1246628572.13604.2.camel@blaa> <1246636154.13604.8.camel@blaa> <1247763853.3038.105.camel@blaa> <1248769817.3089.20.camel@blaa> <1248776732.3089.45.camel@blaa> <1248859649.3021.28.camel@blaa> <20090729093637.GC10723@salstar.sk> <1248868438.3021.31.camel@blaa> <1248873640.3021.34.camel@blaa> <1248879462.3021.39.camel@blaa> <1248882336.3021.40.camel@blaa> <1248974623.3275.4.camel@blaa> <1249057022.3320.35.camel@blaa> <1249400197.3212.73.camel@blaa> <1249573056.3721.62.camel@blaa> <1249639097.3260.5.camel@blaa> <1249900612.8784.79.camel@blaa> <1250700893.17874.44.camel@blaa> <1250703309.17874.46.camel@blaa> Message-ID: <1254834275.2720.47.camel@blaa> The latest updates available from http://markmc.fedorapeople.org/virt-preview == qemu == * Mon Oct 5 2009 Mark McLoughlin - 2:0.11.0-4 - Use rtl8029 PXE rom for ne2k_pci, not ne (#526777) - Also, replace the gpxe-roms-qemu pkg requires with file-based requires * Thu Oct 1 2009 Justin M. Forbes - 2:0.11.0-3 - Improve error reporting on file access (#524695) == libvirt == * Tue Oct 6 2009 Mark McLoughlin - 0.7.1-10 - Create /var/log/libvirt/{lxc,uml} dirs for logrotate - Make libvirt-python dependon on libvirt-client - Sync misc minor changes from upstream spec * Tue Oct 6 2009 Mark McLoughlin - 0.7.1-9 - Change logrotate config to weekly (#526769) == python-virtinst == * Mon Oct 05 2009 Cole Robinson - 0.500.0-5.fc12 - Update translations (bz 493795) == virt-manager == * Mon Oct 05 2009 Cole Robinson - 0.8.0-7.fc12 - More translations (bz 493795) - Don't allow creating a volume without a name (bz 526111) - Don't allow volume allocation > capacity (bz 526077) - Add tooltips for toolbar buttons (bz 524083) Cheers, Mark. == About The Virtualization Preview Repository == The virt-preview repository offers Fedora 11 users the opportunity to help out with testing the rawhide virtualization packages, without having to run all of rawhide. Bugs should be reported to http://bugzilla.redhat.com/ using 'rawhide' as the distribution version. Note clearly in the bug report that you are running these packages on Fedora 11. From markmc at redhat.com Tue Oct 6 13:27:10 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Tue, 06 Oct 2009 14:27:10 +0100 Subject: [fedora-virt] ksmtuned, v2 In-Reply-To: <20090923075235.GB5396@redhat.com> References: <20090903114922.GA11622@redhat.com> <20090903120547.GS8626@redhat.com> <20090904151927.GA11224@csdan.redhat.com> <20090910095055.GD1257@redhat.com> <20090915142244.GA19182@redhat.com> <1253095357.6350.20.camel@blaa> <20090916124007.GA30281@redhat.com> <1253119587.6350.33.camel@blaa> <20090916181900.GA9835@csdan.redhat.com> <20090916182113.GA18548@redhat.com> <20090923075235.GB5396@redhat.com> Message-ID: <1254835630.2720.49.camel@blaa> Hi Dan, Sorry for taking so long to get around to this On Wed, 2009-09-23 at 10:52 +0300, Dan Kenigsberg wrote: > On Wed, Sep 16, 2009 at 07:21:14PM +0100, Daniel P. Berrange wrote: > > On Wed, Sep 16, 2009 at 09:19:01PM +0300, Dan Kenigsberg wrote: > > > On Wed, Sep 16, 2009 at 05:46:27PM +0100, Mark McLoughlin wrote: > > > > > > > > Pushing this to rawhide now > > > > > > Thanks. Though only now did I notice that you dropped my non-standard > > > "signal" verb. http://gitorious.org/ksm-control-scripts/ksm-control-scripts/commit/84e59d1e2d4c243f010c8a4fdd1c4d147e1cd2ee > > > This is how I want managemet to tell ksmtune that something has changed > > > (new qemu process up, or just died). I want ksm to kick in as soon as > > > this happens, not wait another minute. > > > > > > If you really hate this, we can add a SIGUSR handler to ksmtune for the > > > same aim. > > > > I think your 'signal' verb makes sense - I'd just call it 'reload' instead > > > > Mark, would you prefer the attached implementation? (I know I do) > DPB, is the name 'retune' good enough? (I just do not see what is 'loaded') Yeah, that makes sense - e.g. 'retune' doesn't reload ksmtuned.conf I've pushed this now, it'll be in qemu-0.11.0-5.fc12 Thanks, Mark. From markmc at redhat.com Tue Oct 6 13:34:04 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Tue, 06 Oct 2009 14:34:04 +0100 Subject: [fedora-virt] ksmtuned, v2 In-Reply-To: <20090923075235.GB5396@redhat.com> References: <20090903114922.GA11622@redhat.com> <20090903120547.GS8626@redhat.com> <20090904151927.GA11224@csdan.redhat.com> <20090910095055.GD1257@redhat.com> <20090915142244.GA19182@redhat.com> <1253095357.6350.20.camel@blaa> <20090916124007.GA30281@redhat.com> <1253119587.6350.33.camel@blaa> <20090916181900.GA9835@csdan.redhat.com> <20090916182113.GA18548@redhat.com> <20090923075235.GB5396@redhat.com> Message-ID: <1254836044.2720.51.camel@blaa> On Wed, 2009-09-23 at 10:52 +0300, Dan Kenigsberg wrote: > diff --git a/ksmtuned.init b/ksmtuned.init > index 205531a..46332f8 100644 > --- a/ksmtuned.init > +++ b/ksmtuned.init > @@ -75,8 +75,11 @@ case "$1" in > condrestart) > condrestart > ;; > + retune) > + kill -SIGUSR1 `cat ${pidfile}` > + RETVAL=$? > *) > - echo $"Usage: $prog {start|stop|restart|condrestart|status|help}" > + echo $"Usage: $prog {start|stop|restart|condrestart|status|retune|help}" > RETVAL=3 > esac Unusually, I'm actually testing the stuff I'm building today and I hit: /etc/init.d/ksmtuned: line 81: syntax error near unexpected token `)' /etc/init.d/ksmtuned: line 81: ` *)' Missing a ';;' in the handling of retune; fixed now Cheers, Mark. From dennisml at conversis.de Tue Oct 6 14:27:05 2009 From: dennisml at conversis.de (Dennis J.) Date: Tue, 06 Oct 2009 16:27:05 +0200 Subject: [fedora-virt] Providing access to a single VM Message-ID: <4ACB53B9.600@conversis.de> Hi, is there a way to provide virt-viewer access to a single VM? After some googling I can only find ways to provide access to a whole virt-manager instance but I'd like to create VMs for users an only give them access to their own virtual machines? Is this possible? Regards, Dennis From andres at verot.com Wed Oct 7 11:17:06 2009 From: andres at verot.com (=?ISO-8859-1?Q?Andr=E9s_Garc=EDa?=) Date: Wed, 07 Oct 2009 13:17:06 +0200 Subject: [fedora-virt] Re: parameters In-Reply-To: <4AC4E8C1.9020707@verot.com> References: <4AC4E8C1.9020707@verot.com> Message-ID: <4ACC78B2.4050208@verot.com> > > > >
> > > > 'bus' and 'slot' I can get with 'lsusb', but what does 'function' mean? > > > Looking it up in the libvirt sources, it appears to be used only for pci devices. Andr?s From rjones at redhat.com Wed Oct 7 11:24:45 2009 From: rjones at redhat.com (Richard W.M. Jones) Date: Wed, 7 Oct 2009 12:24:45 +0100 Subject: [fedora-virt] Domain doesn't show up In-Reply-To: References: Message-ID: <20091007112445.GA25639@amd.home.annexia.org> On Mon, Oct 05, 2009 at 11:45:09PM +0100, Tim Richardson wrote: > I'm trying to install xp virtual machine from an .iso on the disk, > I have a virtual machine called localhost with qemu as ID. It has a storage > area at /var/lib/libvirt/images, i'm right clicking on that and choosing new > then following the wizard but it keeps failing whilst creating the domain. > Is anyone familiar with this scenario or have any suggestions? > > regards > Tim > > Unable to complete install ' internal error > Domain microXP didn't show up > > Traceback (most recent call last): > File "/usr/share/virt-manager/virtManager/create.py", line 1501, in > do_install > dom = guest.start_install(False, meter = meter) > File "/usr/lib/python2.6/site-packages/virtinst/Guest.py", line 541, in > start_install > return self._do_install(consolecb, meter, removeOld, wait) > File "/usr/lib/python2.6/site-packages/virtinst/Guest.py", line 633, in > _do_install > self.domain = self.conn.createLinux(install_xml, 0) > File "/usr/lib64/python2.6/site-packages/libvirt.py", line 974, in > createLinux > if ret is None:raise libvirtError('virDomainCreateLinux() failed', > conn=self) > libvirtError: internal error Domain microXP didn't show up The above messages don't help. Try looking at the files in /root/.virtmanager/ and/or /var/log/libvirt/ Rich. -- Richard Jones, Emerging Technologies, Red Hat http://et.redhat.com/~rjones Read my programming blog: http://rwmj.wordpress.com Fedora now supports 75 OCaml packages (the OPEN alternative to F#) http://cocan.org/getting_started_with_ocaml_on_red_hat_and_fedora From rjones at redhat.com Wed Oct 7 11:30:15 2009 From: rjones at redhat.com (Richard W.M. Jones) Date: Wed, 7 Oct 2009 12:30:15 +0100 Subject: [fedora-virt] Providing access to a single VM In-Reply-To: <4ACB53B9.600@conversis.de> References: <4ACB53B9.600@conversis.de> Message-ID: <20091007113015.GB25639@amd.home.annexia.org> On Tue, Oct 06, 2009 at 04:27:05PM +0200, Dennis J. wrote: > is there a way to provide virt-viewer access to a single VM? After some > googling I can only find ways to provide access to a whole virt-manager > instance but I'd like to create VMs for users an only give them access to > their own virtual machines? Is this possible? I think you're going beyond what can be done using virt-manager. Have a look at the XML configuration of your domains (virsh dumpxml guestname http://libvirt.org/formatdomain.html). You'll see that you can assign a port number to each virtual machine by using: Give each guest a different port number and control access to those ports (eg. using the firewall or ssh port forwarding). Rich. -- Richard Jones, Emerging Technologies, Red Hat http://et.redhat.com/~rjones Read my programming blog: http://rwmj.wordpress.com Fedora now supports 75 OCaml packages (the OPEN alternative to F#) http://cocan.org/getting_started_with_ocaml_on_red_hat_and_fedora From kashyapc at fedoraproject.org Thu Oct 8 06:07:35 2009 From: kashyapc at fedoraproject.org (Kashyap Chamarthy) Date: Thu, 8 Oct 2009 11:37:35 +0530 Subject: [fedora-virt] question on virt-install (and partly anaconda) Message-ID: <7bbaf0950910072307t6ab353adufcd02ef125745061@mail.gmail.com> Hi, I run something like this, for un-attended automated guest installs. =============================================================================================== #!/bin/bash domname=yellow location=http://10.65.x.x/fedora/linux/releases/11/Fedora/x86_64/os/ vmimage="/var/lib/libvirt/images/$domname.img" echo "Creating domain $domname" echo "Image is here $vmimage" echo "Location of the OS sources $location" virt-install --connect=qemu:///system \ --network=bridge:br0 \ --extra-args="ks=http://10.65.y.y/fedora-minimal.ks console=tty0 console=ttyS0,9600" \ --name $domname \ --file=$vmimage \ --file-size=8 \ --ram 1024 \ --accelerate \ --nonsparse \ --location=$location \ ============================================================================================ now here, -- once the install is finished and guest reboots, a console does /not/ show up. Because, as the "console=tty0 console=ttyS0,9600" args from the script are not picked by the anaconda of guest while booting. -- A console shows up only when I manually edit the guest machine's kernel args and append "console=tty0 console=ttyS0,9600" , /then/ boot the guest - now console shows up happily. My question is : shouldn't the anaconda of the guest machine pick up automatically "console=tty0 console=ttyS0,9600" in the --extra-args mentioned in the virt-install above? Is this the expected behaviour? or shall I file a bug. Regards, Kashyap -------------- next part -------------- An HTML attachment was scrubbed... URL: From clalance at redhat.com Thu Oct 8 07:31:49 2009 From: clalance at redhat.com (Chris Lalancette) Date: Thu, 08 Oct 2009 09:31:49 +0200 Subject: [fedora-virt] question on virt-install (and partly anaconda) In-Reply-To: <7bbaf0950910072307t6ab353adufcd02ef125745061@mail.gmail.com> References: <7bbaf0950910072307t6ab353adufcd02ef125745061@mail.gmail.com> Message-ID: <4ACD9565.7090209@redhat.com> Kashyap Chamarthy wrote: > Hi, > > I run something like this, for un-attended automated guest installs. > =============================================================================================== > #!/bin/bash > > domname=yellow > > location=http://10.65.x.x/fedora/linux/releases/11/Fedora/x86_64/os/ > vmimage="/var/lib/libvirt/images/$domname.img" > > echo "Creating domain $domname" > > echo "Image is here $vmimage" > echo "Location of the OS sources $location" > > virt-install --connect=qemu:///system \ > --network=bridge:br0 \ > --extra-args="ks=http://10.65.y.y/fedora-minimal.ks console=tty0 console=ttyS0,9600" \ > > --name $domname \ > --file=$vmimage \ > --file-size=8 \ > --ram 1024 \ > --accelerate \ > --nonsparse \ > --location=$location \ > ============================================================================================ > > now here, > > -- once the install is finished and guest reboots, a console does /not/ show up. Because, as the > "console=tty0 console=ttyS0,9600" args from the script are not picked by the anaconda of guest while booting. > > > -- A console shows up only when I manually edit the guest machine's kernel args and append > "console=tty0 console=ttyS0,9600" , /then/ boot the guest - now console shows up happily. > > My question is : shouldn't the anaconda of the guest machine pick up automatically > > "console=tty0 console=ttyS0,9600" in the --extra-args mentioned in the virt-install above? > > Is this the expected behaviour? or shall I file a bug. Hm, interesting. I think anaconda used to pick those up. It sounds like a problem in anaconda; I would file a bug about it. For a workaround, you can edit your kickstart with something like: bootloader --location=mbr --append="console=tty0 console=ttyS0,9600" (this certainly works, I use it all the time) -- Chris Lalancette From kashyapc at fedoraproject.org Thu Oct 8 08:48:08 2009 From: kashyapc at fedoraproject.org (Kashyap Chamarthy) Date: Thu, 8 Oct 2009 14:18:08 +0530 Subject: [fedora-virt] question on virt-install (and partly anaconda) In-Reply-To: <4ACD9565.7090209@redhat.com> References: <7bbaf0950910072307t6ab353adufcd02ef125745061@mail.gmail.com> <4ACD9565.7090209@redhat.com> Message-ID: <7bbaf0950910080148pf279f8bocdd459474f1ba73f@mail.gmail.com> On Thu, Oct 8, 2009 at 1:01 PM, Chris Lalancette wrote: > Kashyap Chamarthy wrote: > > Hi, > > > > I run something like this, for un-attended automated guest installs. > > > =============================================================================================== > > #!/bin/bash > > > > domname=yellow > > > > location=http://10.65.x.x/fedora/linux/releases/11/Fedora/x86_64/os/ > > vmimage="/var/lib/libvirt/images/$domname.img" > > > > echo "Creating domain $domname" > > > > echo "Image is here $vmimage" > > echo "Location of the OS sources $location" > > > > virt-install --connect=qemu:///system \ > > --network=bridge:br0 \ > > --extra-args="ks=http://10.65.y.y/fedora-minimal.ks console=tty0 > console=ttyS0,9600" \ > > > > --name $domname \ > > --file=$vmimage \ > > --file-size=8 \ > > --ram 1024 \ > > --accelerate \ > > --nonsparse \ > > --location=$location \ > > > ============================================================================================ > > > > now here, > > > > -- once the install is finished and guest reboots, a console does /not/ > show up. Because, as the > > "console=tty0 console=ttyS0,9600" args from the script are not picked by > the anaconda of guest while booting. > > > > > > -- A console shows up only when I manually edit the guest machine's > kernel args and append > > "console=tty0 console=ttyS0,9600" , /then/ boot the guest - now console > shows up happily. > > > > My question is : shouldn't the anaconda of the guest machine pick up > automatically > > > > "console=tty0 console=ttyS0,9600" in the --extra-args mentioned in the > virt-install above? > > > > Is this the expected behaviour? or shall I file a bug. > > Hm, interesting. I think anaconda used to pick those up. It sounds like a > problem in anaconda; I would file a bug about it. For a workaround, you > can > edit your kickstart with something like: > > bootloader --location=mbr --append="console=tty0 console=ttyS0,9600" > > (this certainly works, I use it all the time) > right, If I edit the kickstart file, it'll pick. thanks for confirming. (so I guess you're filing the bug. Ifyou want me to I can do that.) -- Kashyap Chamarthy > > -- > Chris Lalancette > -------------- next part -------------- An HTML attachment was scrubbed... URL: From clalance at redhat.com Thu Oct 8 08:58:13 2009 From: clalance at redhat.com (Chris Lalancette) Date: Thu, 08 Oct 2009 10:58:13 +0200 Subject: [fedora-virt] question on virt-install (and partly anaconda) In-Reply-To: <7bbaf0950910080148pf279f8bocdd459474f1ba73f@mail.gmail.com> References: <7bbaf0950910072307t6ab353adufcd02ef125745061@mail.gmail.com> <4ACD9565.7090209@redhat.com> <7bbaf0950910080148pf279f8bocdd459474f1ba73f@mail.gmail.com> Message-ID: <4ACDA9A5.1090804@redhat.com> Kashyap Chamarthy wrote: > right, If I edit the kickstart file, it'll pick. thanks for confirming. > (so I guess you're filing the bug. Ifyou want me to I can do that.) I would prefer you open the bug, since I don't have personal knowledge of the bug (and thus can't respond to requests for more information). -- Chris Lalancette From pavel.lisy at gmail.com Thu Oct 8 11:09:35 2009 From: pavel.lisy at gmail.com (Pavel Lisy) Date: Thu, 08 Oct 2009 13:09:35 +0200 Subject: [fedora-virt] libvirtd and public access to guests Message-ID: <1255000175.20724.34.camel@pali-pc.hk.tmapy.cz> Hello I've started playing with libvirt and I have question? What is proper way to make guest accessible from net. I have mode=nat /var/lib/libvirt/network/default.xml. libvirtd makes this rules in FORWARD chain -A FORWARD -d 192.168.231.0/24 -o virbr0 -m state --state RELATED,ESTABLISHED -j ACCEPT -A FORWARD -s 192.168.231.0/24 -i virbr0 -j ACCEPT -A FORWARD -i virbr0 -o virbr0 -j ACCEPT -A FORWARD -o virbr0 -j REJECT --reject-with icmp-port-unreachable -A FORWARD -i virbr0 -j REJECT --reject-with icmp-port-unreachable -A FORWARD -j REJECT --reject-with icmp-host-prohibited If I add iptables -I FORWARD -i eth0 -o virbr0 -j ACCEPT guests are accessible My question is: Is is possible write this somewhere to configuration? I've tried to put it in /etc/sysconfig/iptables but it libvirtd put his rules before mine. I've found two directories /var/lib/libvirt/iptables/filter /var/lib/libvirt/iptables/nat I suppose I can write my rules here but I haven't find any docs about format. Can somebody help me with it? Pavel From rjones at redhat.com Thu Oct 8 14:09:25 2009 From: rjones at redhat.com (Richard W.M. Jones) Date: Thu, 8 Oct 2009 15:09:25 +0100 Subject: [fedora-virt] libvirtd and public access to guests In-Reply-To: <1255000175.20724.34.camel@pali-pc.hk.tmapy.cz> References: <1255000175.20724.34.camel@pali-pc.hk.tmapy.cz> Message-ID: <20091008140651.GA8507@amd.home.annexia.org> On Thu, Oct 08, 2009 at 01:09:35PM +0200, Pavel Lisy wrote: > I've started playing with libvirt and I have question? This question is probably better asked on libvir-list. https://www.redhat.com/mailman/listinfo/libvir-list > What is proper way to make guest accessible from net. > > I have mode=nat /var/lib/libvirt/network/default.xml. > > libvirtd makes this rules in FORWARD chain > > -A FORWARD -d 192.168.231.0/24 -o virbr0 -m state --state > RELATED,ESTABLISHED -j ACCEPT > -A FORWARD -s 192.168.231.0/24 -i virbr0 -j ACCEPT > -A FORWARD -i virbr0 -o virbr0 -j ACCEPT > -A FORWARD -o virbr0 -j REJECT --reject-with icmp-port-unreachable > -A FORWARD -i virbr0 -j REJECT --reject-with icmp-port-unreachable > -A FORWARD -j REJECT --reject-with icmp-host-prohibited > > If I add > iptables -I FORWARD -i eth0 -o virbr0 -j ACCEPT > guests are accessible An alternate way is to create your own bridge (however you want to configure it), then make it a network that guests can see and connect to, using commands like 'virsh net-create', 'virsh net-dumpxml' and 'virsh net-edit'. The XML format is described here: http://libvirt.org/formatnetwork.html > My question is: > Is is possible write this somewhere to configuration? > > I've tried to put it in /etc/sysconfig/iptables but it libvirtd put his > rules before mine. I think libvirtd will trash your virbr0 definitions, so maybe setting up your own bridge is a better idea. Rich. -- Richard Jones, Emerging Technologies, Red Hat http://et.redhat.com/~rjones virt-df lists disk usage of guests without needing to install any software inside the virtual machine. Supports Linux and Windows. http://et.redhat.com/~rjones/virt-df/ From jlaska at redhat.com Thu Oct 8 15:44:38 2009 From: jlaska at redhat.com (James Laska) Date: Thu, 08 Oct 2009 11:44:38 -0400 Subject: [fedora-virt] Should python-virtinst %require libvirt? Message-ID: <1255016678.2376.10.camel@localhost.localdomain> Greetings, Quick question about virt package dependencies. The Rawhide Acceptance Install Test (which does a virt install) [1] first installs the required packages by doing: # yum -y install qemu-kvm python-virtinst pax It seems that installing python-virtinst used to also pull libvirt into the transaction. However, this is no longer the case [2]. Should the deps chain when installing python-virtinst also install libvirt? Or is it expected that if you wish to run your own virt daemon, you should install libvirtd? Thanks, James [1] http://git.fedorahosted.org/git/?p=autoqa.git;a=blob;f=tests/rats_install/rats_install.py;h=950d1538b443d78a6f89f39e07edb6b013cb2906;hb=HEAD [2] rats_install test output - http://pastie.org/647021 -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 198 bytes Desc: This is a digitally signed message part URL: From berrange at redhat.com Thu Oct 8 15:49:17 2009 From: berrange at redhat.com (Daniel P. Berrange) Date: Thu, 8 Oct 2009 16:49:17 +0100 Subject: [fedora-virt] Should python-virtinst %require libvirt? In-Reply-To: <1255016678.2376.10.camel@localhost.localdomain> References: <1255016678.2376.10.camel@localhost.localdomain> Message-ID: <20091008154917.GN17867@redhat.com> On Thu, Oct 08, 2009 at 11:44:38AM -0400, James Laska wrote: > Greetings, > > Quick question about virt package dependencies. The Rawhide Acceptance > Install Test (which does a virt install) [1] first installs the required > packages by doing: > > # yum -y install qemu-kvm python-virtinst pax > > It seems that installing python-virtinst used to also pull libvirt into > the transaction. However, this is no longer the case [2]. Should the > deps chain when installing python-virtinst also install libvirt? Or is > it expected that if you wish to run your own virt daemon, you should > install libvirtd? In previous Fedora: python-virtinst -> libvirt-python libvirt-python -> libvirt (and thus libvirtd) In F12 though we split libvirt python-virtinst -> libvirt-python livirt-python -> libvirt-client So, python-virtinst no longer causes the libvirtd daemon to be pulled in In the general case this is good, because we explicitly want to allow a client only install of virtualization tools. If wanting to actually setup a virtualization host though, your yum command line is no longer sufficient. You probably want to change to use yum -y groupinstall Virtualization Which I believe should pull in KVM, libvirt, virt-manager & virtinst Daniel -- |: Red Hat, Engineering, London -o- http://people.redhat.com/berrange/ :| |: http://libvirt.org -o- http://virt-manager.org -o- http://ovirt.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: GnuPG: 7D3B9505 -o- F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 :| From jlaska at redhat.com Thu Oct 8 17:58:49 2009 From: jlaska at redhat.com (James Laska) Date: Thu, 08 Oct 2009 13:58:49 -0400 Subject: [fedora-virt] Should python-virtinst %require libvirt? In-Reply-To: <20091008154917.GN17867@redhat.com> References: <1255016678.2376.10.camel@localhost.localdomain> <20091008154917.GN17867@redhat.com> Message-ID: <1255024729.2376.15.camel@localhost.localdomain> On Thu, 2009-10-08 at 16:49 +0100, Daniel P. Berrange wrote: > On Thu, Oct 08, 2009 at 11:44:38AM -0400, James Laska wrote: > > Greetings, > > > > Quick question about virt package dependencies. The Rawhide Acceptance > > Install Test (which does a virt install) [1] first installs the required > > packages by doing: > > > > # yum -y install qemu-kvm python-virtinst pax > > > > It seems that installing python-virtinst used to also pull libvirt into > > the transaction. However, this is no longer the case [2]. Should the > > deps chain when installing python-virtinst also install libvirt? Or is > > it expected that if you wish to run your own virt daemon, you should > > install libvirtd? > > In previous Fedora: > > python-virtinst -> libvirt-python > libvirt-python -> libvirt (and thus libvirtd) > > In F12 though we split libvirt > > python-virtinst -> libvirt-python > livirt-python -> libvirt-client > > > So, python-virtinst no longer causes the libvirtd daemon to be pulled in > > In the general case this is good, because we explicitly want to allow a > client only install of virtualization tools. > > If wanting to actually setup a virtualization host though, your yum > command line is no longer sufficient. > > You probably want to change to use > > yum -y groupinstall Virtualization > > Which I believe should pull in KVM, libvirt, virt-manager & virtinst Thanks for the explanation Daniel! That helps. -James -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 198 bytes Desc: This is a digitally signed message part URL: From berrange at redhat.com Thu Oct 8 20:57:01 2009 From: berrange at redhat.com (Daniel P. Berrange) Date: Thu, 8 Oct 2009 21:57:01 +0100 Subject: [fedora-virt] libvirtd and public access to guests In-Reply-To: <1255000175.20724.34.camel@pali-pc.hk.tmapy.cz> References: <1255000175.20724.34.camel@pali-pc.hk.tmapy.cz> Message-ID: <20091008205701.GB23950@redhat.com> On Thu, Oct 08, 2009 at 01:09:35PM +0200, Pavel Lisy wrote: > Hello > > I've started playing with libvirt and I have question? > > What is proper way to make guest accessible from net. The shared physical device, bridging option is what you want to use http://wiki.libvirt.org/page/Networking#Fedora.2FRHEL_Bridging > I have mode=nat /var/lib/libvirt/network/default.xml. NAT is for outbound internet access only - it doesn't allow for remote clients to connect to your VM. > libvirtd makes this rules in FORWARD chain > > -A FORWARD -d 192.168.231.0/24 -o virbr0 -m state --state > RELATED,ESTABLISHED -j ACCEPT > -A FORWARD -s 192.168.231.0/24 -i virbr0 -j ACCEPT > -A FORWARD -i virbr0 -o virbr0 -j ACCEPT > -A FORWARD -o virbr0 -j REJECT --reject-with icmp-port-unreachable > -A FORWARD -i virbr0 -j REJECT --reject-with icmp-port-unreachable > -A FORWARD -j REJECT --reject-with icmp-host-prohibited > > If I add > iptables -I FORWARD -i eth0 -o virbr0 -j ACCEPT > guests are accessible > > My question is: > Is is possible write this somewhere to configuration? > > I've tried to put it in /etc/sysconfig/iptables but it libvirtd put his > rules before mine. > > I've found two directories > /var/lib/libvirt/iptables/filter > /var/lib/libvirt/iptables/nat > > I suppose I can write my rules here but I haven't find any docs about > format. Can somebody help me with it? You shouldn't try to overwrite/override libvirt's rules here, since libvirt will likely just break your changes at some point. You really want to switch to a bridged network config, instead of the NAT based one Daniel -- |: Red Hat, Engineering, London -o- http://people.redhat.com/berrange/ :| |: http://libvirt.org -o- http://virt-manager.org -o- http://ovirt.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: GnuPG: 7D3B9505 -o- F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 :| From berrange at redhat.com Thu Oct 8 20:58:36 2009 From: berrange at redhat.com (Daniel P. Berrange) Date: Thu, 8 Oct 2009 21:58:36 +0100 Subject: [fedora-virt] Re: parameters In-Reply-To: <4ACC78B2.4050208@verot.com> References: <4AC4E8C1.9020707@verot.com> <4ACC78B2.4050208@verot.com> Message-ID: <20091008205836.GC23950@redhat.com> On Wed, Oct 07, 2009 at 01:17:06PM +0200, Andr?s Garc?a wrote: > > > > > > > > >
> > > > > > > >'bus' and 'slot' I can get with 'lsusb', but what does 'function' mean? > > > > > > > Looking it up in the libvirt sources, it appears to be used only for pci > devices. For USB passthrough, you want
The bus/slot/domain/function style is for PCI device addressing. Daniel -- |: Red Hat, Engineering, London -o- http://people.redhat.com/berrange/ :| |: http://libvirt.org -o- http://virt-manager.org -o- http://ovirt.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: GnuPG: 7D3B9505 -o- F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 :| From rich at lat.com Fri Oct 9 13:45:36 2009 From: rich at lat.com (Rich Mahn) Date: Fri, 09 Oct 2009 09:45:36 -0400 Subject: [fedora-virt] libvirtd crashes host when autostarting Message-ID: <200910091345.n99DjaB0006654@raspberry.lat.com> Summary: My host crashes when I use the autostart function in libvirtd. This command sequence works fine: service libvirtd stop virsh autostart mydomain --disable service lib virtd start virsh start mydomain This command sequence will crash the host: service libvirtd stop virsh autostart mydomain service libvirtd start I can't find any error messages related to this. The host system freezes, the keyboard starts blinking the caps-lock and scroll-lock, and as best I can tell, the kernel has stopped running. If I starting up the host system with the VM autostarted, I sometimes get messages that look like kernel errors on the screen, but I don't find them in any log files so far. They seem to be related to networking. Here's some configuration information--I don't know really what all is needed, but this, I guess is a start: Host and VM both Fedora 11 with the lastest updates. Host machine in Intel based with the virtual support turned on. In the /var/log/libvirtd/qemu directory, the command line to start the virtual machine is different when started from autostart. The difference is in this portion: -net tap,fd=11,vlan=0 for failure (autostart) and -net tap,fd=18,vlan=0 for success I am using a bridged connection. Here's the command line that's used, split over lines for some degree of readability: LC_ALL=C PATH=/sbin:/usr/sbin:/bin:/usr/bin /usr/bin/qemu-kvm -S -M pc -m 1024 -smp 1 -name mydomain -uuid 5cde5a23-2abe-fcdd-04dc-5ae3a46ab51c -monitor pty -pidfile /var/run/libvirt/qemu//mydomain.pid -boot c -drive file=,if=ide,media=cdrom,index=2 -drive file=/dev/mapper/vg_host-lv_mydomain,if=virtio,index=0,boot=on -drive file=/dev/mapper/vg_host-lv_mydomain_data,if=virtio,index=1 -net nic,macaddr=54:52:00:67:91:e9,vlan=0,model=virtio -net tap,fd=11,vlan=0 -serial pty -parallel none -usb -usbdevice tablet -vnc 127.0.0.1:0 From markmc at redhat.com Fri Oct 9 13:48:35 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Fri, 09 Oct 2009 14:48:35 +0100 Subject: [fedora-virt] user qemu has no access to usb devices In-Reply-To: <100320091938.13024.4AC7A82F000BA317000032E022216125569B0A02D29B9B0EBF970A049C9D0108D203019B@att.net> References: <100320091938.13024.4AC7A82F000BA317000032E022216125569B0A02D29B9B0EBF970A049C9D0108D203019B@att.net> Message-ID: <1255096115.2786.58.camel@blaa> On Sat, 2009-10-03 at 19:38 +0000, Tom Horsley wrote: > I don't know if something is busted in my rawhide setup, or if this is just > a real problem with the new scheme of running as user qemu, but user qemu (on > my system at least), has no access to physical usb devices. I had to change > the qemu.conf file to tell it to run as root before it could attach a usb > device to my windows XP virtual machine. > > See https://bugzilla.redhat.com/show_bug.cgi?id=524723#c5 Just for the record, a better workaround is to enable the virt_use_usb boolean with 'setsebool virt_use_usb true' Cheers, Mark. From markmc at redhat.com Fri Oct 9 13:53:47 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Fri, 09 Oct 2009 14:53:47 +0100 Subject: [fedora-virt] libvirtd crashes host when autostarting In-Reply-To: <200910091345.n99DjaB0006654@raspberry.lat.com> References: <200910091345.n99DjaB0006654@raspberry.lat.com> Message-ID: <1255096427.2786.61.camel@blaa> On Fri, 2009-10-09 at 09:45 -0400, Rich Mahn wrote: > Summary: > My host crashes when I use the autostart function in > libvirtd. > > This command sequence works fine: > > service libvirtd stop > virsh autostart mydomain --disable > service lib virtd start > virsh start mydomain > > This command sequence will crash the host: > > service libvirtd stop > virsh autostart mydomain > service libvirtd start > > > I can't find any error messages related to this. The host system > freezes, the keyboard starts blinking the caps-lock and scroll-lock, > and as best I can tell, the kernel has stopped running. Ouch, that's really bad. This is a 2.6.30 kernel, right? Is it still reproducible with an older 2.6.29 kernel or a 2.6.31 kernel from rawhide? This is ultimately a kernel bug, userspace shouldn't be able to screw up the kernel like this. It'd be good to get this into bugzilla: https://fedoraproject.org/wiki/Reporting virtualization bugs > If I starting up the host system with the VM autostarted, I sometimes > get messages that look like kernel errors on the screen, but I don't > find them in any log files so far. They seem to be related to networking. > > Here's some configuration information--I don't know really what all is > needed, but this, I guess is a start: > > Host and VM both Fedora 11 with the lastest updates. > Host machine in Intel based with the virtual support turned on. > > In the /var/log/libvirtd/qemu directory, the command line to start > the virtual machine is different when started from autostart. The > difference is in this portion: > > -net tap,fd=11,vlan=0 for failure (autostart) and > -net tap,fd=18,vlan=0 for success Interesting, but I can't immediately think where the difference may come from or how it could be related to the kernel lockup. Cheers, Mark. From eb30750 at gmail.com Fri Oct 9 14:18:09 2009 From: eb30750 at gmail.com (Paul Lambert) Date: Fri, 9 Oct 2009 10:18:09 -0400 Subject: [fedora-virt] libvirtd crashes host when autostarting In-Reply-To: <1255096427.2786.61.camel@blaa> References: <200910091345.n99DjaB0006654@raspberry.lat.com> <1255096427.2786.61.camel@blaa> Message-ID: Rich, Mark, I experienced this problem as well when using a FE-11 host and FE-11 Guest. I now have FE-12-alpha installed but have been afraid to try it until I know if this issue was addressed in the new version. Configuring a guest to start on host bootup totally hoses the host. The only way to recover is to boot off of a CD and then find the startup directory and delete the link to the guest. The host will reboot OK after this. By the way Mark, still no mounting for USB disks with FE-12-Alpha. Up to date on patches as of 10/8/2009. Paul On Fri, Oct 9, 2009 at 9:53 AM, Mark McLoughlin wrote: > On Fri, 2009-10-09 at 09:45 -0400, Rich Mahn wrote: > > Summary: > > My host crashes when I use the autostart function in > > libvirtd. > > > > This command sequence works fine: > > > > service libvirtd stop > > virsh autostart mydomain --disable > > service lib virtd start > > virsh start mydomain > > > > This command sequence will crash the host: > > > > service libvirtd stop > > virsh autostart mydomain > > service libvirtd start > > > > > > I can't find any error messages related to this. The host system > > freezes, the keyboard starts blinking the caps-lock and scroll-lock, > > and as best I can tell, the kernel has stopped running. > > Ouch, that's really bad. This is a 2.6.30 kernel, right? Is it still > reproducible with an older 2.6.29 kernel or a 2.6.31 kernel from > rawhide? > > This is ultimately a kernel bug, userspace shouldn't be able to screw up > the kernel like this. > > It'd be good to get this into bugzilla: > > https://fedoraproject.org/wiki/Reporting virtualization bugs > > > If I starting up the host system with the VM autostarted, I sometimes > > get messages that look like kernel errors on the screen, but I don't > > find them in any log files so far. They seem to be related to > networking. > > > > Here's some configuration information--I don't know really what all is > > needed, but this, I guess is a start: > > > > Host and VM both Fedora 11 with the lastest updates. > > Host machine in Intel based with the virtual support turned on. > > > > In the /var/log/libvirtd/qemu directory, the command line to start > > the virtual machine is different when started from autostart. The > > difference is in this portion: > > > > -net tap,fd=11,vlan=0 for failure (autostart) and > > -net tap,fd=18,vlan=0 for success > > Interesting, but I can't immediately think where the difference may come > from or how it could be related to the kernel lockup. > > Cheers, > Mark. > > _______________________________________________ > Fedora-virt mailing list > Fedora-virt at redhat.com > https://www.redhat.com/mailman/listinfo/fedora-virt > -------------- next part -------------- An HTML attachment was scrubbed... URL: From markmc at redhat.com Fri Oct 9 14:20:08 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Fri, 09 Oct 2009 15:20:08 +0100 Subject: [fedora-virt] libvirtd crashes host when autostarting In-Reply-To: References: <200910091345.n99DjaB0006654@raspberry.lat.com> <1255096427.2786.61.camel@blaa> Message-ID: <1255098008.2786.63.camel@blaa> On Fri, 2009-10-09 at 10:18 -0400, Paul Lambert wrote: > Rich, Mark, > > I experienced this problem as well when using a FE-11 host and FE-11 > Guest. I now have FE-12-alpha installed but have been afraid to try > it until I know if this issue was addressed in the new version. > > Configuring a guest to start on host bootup totally hoses the host. > The only way to recover is to boot off of a CD and then find the > startup directory and delete the link to the guest. The host will > reboot OK after this. Ouch. We really need to get to the bottom of this then. > By the way Mark, still no mounting for USB disks with FE-12-Alpha. Up > to date on patches as of 10/8/2009. Have you updated to selinux-policy-3.6.32-17.fc12.noarch ? Do you have the virt_use_usb sebool enabled? Thanks, Mark. From rich at lat.com Fri Oct 9 14:23:23 2009 From: rich at lat.com (Rich Mahn) Date: Fri, 09 Oct 2009 10:23:23 -0400 Subject: [fedora-virt] libvirtd crashes host when autostarting In-Reply-To: Your message of "Fri, 09 Oct 2009 10:18:09 EDT." Message-ID: <200910091423.n99ENNK4008652@raspberry.lat.com> > Rich, Mark, > I experienced this problem as well when using a FE-11 host and FE-11 Guest. > I now have FE-12-alpha installed but have been afraid to try it until I know > if this issue was addressed in the new version. > Configuring a guest to start on host bootup totally hoses the host. The > only way to recover is to boot off of a CD and then find the startup > directory and delete the link to the guest. The host will reboot OK after > this. Paul, If you are using grub, type 'a' on the kernel selection, replace 'rhbg quiet' at the end of the line with 'single' ahd boot. Then go ahead and remove the link. Much easier than booting from CD. rich From rich at lat.com Fri Oct 9 14:39:37 2009 From: rich at lat.com (Rich Mahn) Date: Fri, 09 Oct 2009 10:39:37 -0400 Subject: [fedora-virt] libvirtd crashes host when autostarting In-Reply-To: Your message of "Fri, 09 Oct 2009 14:53:47 BST." <1255096427.2786.61.camel@blaa> Message-ID: <200910091439.n99EdbJr009341@raspberry.lat.com> > On Fri, 2009-10-09 at 09:45 -0400, Rich Mahn wrote: > > Summary: > > My host crashes when I use the autostart function in > > libvirtd. > > > > This command sequence works fine: > > > > service libvirtd stop > > virsh autostart mydomain --disable > > service lib virtd start > > virsh start mydomain > > > > This command sequence will crash the host: > > > > service libvirtd stop > > virsh autostart mydomain > > service libvirtd start > > > > > > I can't find any error messages related to this. The host system > > freezes, the keyboard starts blinking the caps-lock and scroll-lock, > > and as best I can tell, the kernel has stopped running. > Ouch, that's really bad. This is a 2.6.30 kernel, right? Is it still > reproducible with an older 2.6.29 kernel or a 2.6.31 kernel from > rawhide? okay, it works correctly on 2.6.29. I'm trying to figure out how to install 2.6.31 from rawhide without having dozens/hundreds of other packages modified as well. Is there some option to 'yum' that will do what I need? rich From pavel.lisy at gmail.com Fri Oct 9 15:35:27 2009 From: pavel.lisy at gmail.com (Pavel Lisy) Date: Fri, 09 Oct 2009 17:35:27 +0200 Subject: [fedora-virt] libvirtd and public access to guests In-Reply-To: <20091008205701.GB23950@redhat.com> References: <1255000175.20724.34.camel@pali-pc.hk.tmapy.cz> <20091008205701.GB23950@redhat.com> Message-ID: <1255102527.24821.29.camel@pali-pc.hk.tmapy.cz> Daniel P. Berrange p??e v ?t 08. 10. 2009 v 21:57 +0100: > On Thu, Oct 08, 2009 at 01:09:35PM +0200, Pavel Lisy wrote: > > Hello > > > > I've started playing with libvirt and I have question? > > > > What is proper way to make guest accessible from net. > > The shared physical device, bridging option is what you want > to use > > http://wiki.libvirt.org/page/Networking#Fedora.2FRHEL_Bridging > > > I have mode=nat /var/lib/libvirt/network/default.xml. > > NAT is for outbound internet access only - it doesn't allow > for remote clients to connect to your VM. > > libvirtd makes this rules in FORWARD chain > > > > -A FORWARD -d 192.168.231.0/24 -o virbr0 -m state --state > > RELATED,ESTABLISHED -j ACCEPT > > -A FORWARD -s 192.168.231.0/24 -i virbr0 -j ACCEPT > > -A FORWARD -i virbr0 -o virbr0 -j ACCEPT > > -A FORWARD -o virbr0 -j REJECT --reject-with icmp-port-unreachable > > -A FORWARD -i virbr0 -j REJECT --reject-with icmp-port-unreachable > > -A FORWARD -j REJECT --reject-with icmp-host-prohibited > > > > If I add > > iptables -I FORWARD -i eth0 -o virbr0 -j ACCEPT > > guests are accessible > > > > My question is: > > Is is possible write this somewhere to configuration? > > > > I've tried to put it in /etc/sysconfig/iptables but it libvirtd put his > > rules before mine. > > > > I've found two directories > > /var/lib/libvirt/iptables/filter > > /var/lib/libvirt/iptables/nat > > > > I suppose I can write my rules here but I haven't find any docs about > > format. Can somebody help me with it? > > You shouldn't try to overwrite/override libvirt's rules here, since libvirt > will likely just break your changes at some point. You really want to switch > to a bridged network config, instead of the NAT based one I've tried it but it isn't what I want. I don't want to have guests in the our LAN network. I want to test LDAP replication and samba config for two different offices so I want to make separated networks accessible from our LAN. I've tried routed network before. I've made necessary changes in configuration on our router. routed fe53ef22-ae5b-47c6-ba24-fe21ea3e06a3 It was working but libvirt couldn't give IP addresses to guests through dhcpmasq. Is it normal? Questions: 1. Can I make network in mode='route' and use dhcp for guests? How? 2. NAT is working perfectly to my needs - routing, dhcp, ... (see my config below) Is it possible put extra iptables rules to libvirt configuration? With dhcp worked but I had to change few IP tables rules to make this net accessible: in file default.xml: default fe53ef22-ae5b-47c6-ba24-fe21ea3e06a3 iptables changes: # remove masquerading iptables -t nat -D POSTROUTING -s 192.168.231.0/24 ! -d 192.168.231.0/24 -j MASQUERADE # open virtnet from eth0 iptables -I FORWARD -i eth0 -o virbr0 -j ACCEPT iptables -D FORWARD -d 192.168.231.0/24 -o virbr0 -m state --state RELATED,ESTABLISHED -j ACCEPT Pavel From markmc at redhat.com Fri Oct 9 15:56:32 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Fri, 09 Oct 2009 16:56:32 +0100 Subject: [fedora-virt] virt-preview updates In-Reply-To: <1250703309.17874.46.camel@blaa> References: <1244034684.5001.134.camel@blaa> <1244046483.5001.180.camel@blaa> <1244197459.27876.0.camel@blaa> <1246303956.11688.148.camel@blaa> <1246628572.13604.2.camel@blaa> <1246636154.13604.8.camel@blaa> <1247763853.3038.105.camel@blaa> <1248769817.3089.20.camel@blaa> <1248776732.3089.45.camel@blaa> <1248859649.3021.28.camel@blaa> <20090729093637.GC10723@salstar.sk> <1248868438.3021.31.camel@blaa> <1248873640.3021.34.camel@blaa> <1248879462.3021.39.camel@blaa> <1248882336.3021.40.camel@blaa> <1248974623.3275.4.camel@blaa> <1249057022.3320.35.camel@blaa> <1249400197.3212.73.camel@blaa> <1249573056.3721.62.camel@blaa> <1249639097.3260.5.camel@blaa> <1249900612.8784.79.camel@blaa> <1250700893.17874.44.camel@blaa> <1250703309.17874.46.camel@blaa> Message-ID: <1255103793.2786.70.camel@blaa> The latest updates available from http://markmc.fedorapeople.org/virt-preview == qemu == * Fri Oct 9 2009 Mark McLoughlin - 2:0.11.0-6 - Fix fs errors with virtio and qcow2 backing file (#524734) - Fix ksm initscript errors on kernel missing ksm (#527653) - Add missing Requires(post): getent, useradd, groupadd (#527087) * Tue Oct 6 2009 Mark McLoughlin - 2:0.11.0-5 - Add 'retune' verb to ksmtuned init script == libvirt == * Fri Oct 9 2009 Mark McLoughlin - 0.7.1-11 - Fix libvirtd memory leak during error reply sending (#528162) - Add several PCI hot-unplug typo fixes from upstream Cheers, Mark. == About The Virtualization Preview Repository == The virt-preview repository offers Fedora 11 users the opportunity to help out with testing the rawhide virtualization packages, without having to run all of rawhide. Bugs should be reported to http://bugzilla.redhat.com/ using 'rawhide' as the distribution version. Note clearly in the bug report that you are running these packages on Fedora 11. From rjones at redhat.com Mon Oct 12 08:19:30 2009 From: rjones at redhat.com (Richard W.M. Jones) Date: Mon, 12 Oct 2009 09:19:30 +0100 Subject: [fedora-virt] Re: Question about segmentation fault under febootstrap In-Reply-To: References: Message-ID: <20091012081930.GC15371@amd.home.annexia.org> On Sat, Oct 10, 2009 at 04:56:35PM +0300, Vitaliy Yermolenko wrote: > I'm a big fun of Red Hat/Fedora distributives, and, have tried your > febootstrap Fedora bootstrapping system, which seems very interesting for me > in order to make own lightweighted Fedora-based Linux distributive. > I'm currently using Fedora 11 with fresh 2.6.31.1 kernel on AMD Phenom II X4 > 945 CPU. But, unfortunately, I have got many segmentation faults during > running ./minimal-filesystem.sh - see below. You should ask questions on one of the proper mailing lists. I've CC'd this to fedora-virt. Segfaults are almost certainly because you're trying to install a guest >= F12 on an F11 host. There's some ABI incompatibility in glibc which causes fakeroot to segfault in this case. Install an F11 guest if you have an F11 host, or upgrade to F12 host. Rich. -- Richard Jones, Emerging Technologies, Red Hat http://et.redhat.com/~rjones libguestfs lets you edit virtual machines. Supports shell scripting, bindings from many languages. http://et.redhat.com/~rjones/libguestfs/ See what it can do: http://et.redhat.com/~rjones/libguestfs/recipes.html From ewan at macmahon.me.uk Mon Oct 12 18:52:46 2009 From: ewan at macmahon.me.uk (Ewan Mac Mahon) Date: Mon, 12 Oct 2009 19:52:46 +0100 Subject: [fedora-virt] gPXE and localboot Message-ID: <20091012185245.GE31560@macmahon.me.uk> Hi, According to https://fedoraproject.org/wiki/Features/VirtgPXE the switch to gPXE should make PXE 'localboot' (which etherboot didn't support) start working. I've tried setting up a network booting VM on a host running F11 + virt-preview and while it uses gPXE, loading a pxelinux configuration that tries to do localboot doesn't work - gPXE loops and does DHCP again[1]. The same config does work on real hardware, so am I missing a detail somewhere, do I need a full F12 install, or is this not supposed to be working yet? Ewan [1] screenshot here: http://www-pnp.physics.ox.ac.uk/~macmahon/gPXElocalboot.png -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 197 bytes Desc: not available URL: From markmc at redhat.com Tue Oct 13 06:42:49 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Tue, 13 Oct 2009 07:42:49 +0100 Subject: [fedora-virt] gPXE and localboot In-Reply-To: <20091012185245.GE31560@macmahon.me.uk> References: <20091012185245.GE31560@macmahon.me.uk> Message-ID: <1255416169.3136.12.camel@blaa> Hi Ewan, On Mon, 2009-10-12 at 19:52 +0100, Ewan Mac Mahon wrote: > Hi, > > According to https://fedoraproject.org/wiki/Features/VirtgPXE the switch > to gPXE should make PXE 'localboot' (which etherboot didn't support) > start working. I've tried setting up a network booting VM on a host > running F11 + virt-preview and while it uses gPXE, loading a pxelinux > configuration that tries to do localboot doesn't work - gPXE loops and > does DHCP again[1]. The same config does work on real hardware, so am I > missing a detail somewhere, do I need a full F12 install, or is this not > supposed to be working yet? Thanks for the report, I think it's probably just a bug. You should follow up on: https://bugzilla.redhat.com/472236 Thanks, Mark. From gianluca.cecchi at gmail.com Tue Oct 13 13:29:06 2009 From: gianluca.cecchi at gmail.com (Gianluca Cecchi) Date: Tue, 13 Oct 2009 15:29:06 +0200 Subject: [fedora-virt] how to attach to qemu monitor Message-ID: <561c252c0910130629g3132198qe37e6647a91b6fb0@mail.gmail.com> I don't remember if it is possible to connect to qemu monitor with virt-manager/libvirtd managed guests. If so, how can I accomplish this? Thanks in advance, Gianluca -------------- next part -------------- An HTML attachment was scrubbed... URL: From gianluca.cecchi at gmail.com Tue Oct 13 15:09:17 2009 From: gianluca.cecchi at gmail.com (Gianluca Cecchi) Date: Tue, 13 Oct 2009 17:09:17 +0200 Subject: [fedora-virt] how to attach to qemu monitor In-Reply-To: <4AD4862D.6010009@redhat.com> References: <561c252c0910130629g3132198qe37e6647a91b6fb0@mail.gmail.com> <4AD4862D.6010009@redhat.com> Message-ID: <561c252c0910130809n1b4f7377k85838a09084fe17f@mail.gmail.com> On Tue, Oct 13, 2009 at 3:52 PM, Don Dutile wrote: > Gianluca Cecchi wrote: > >> I don't remember if it is possible to connect to qemu monitor with >> virt-manager/libvirtd managed guests. >> If so, how can I accomplish this? >> >> Thanks in advance, >> Gianluca >> > do you mean?: > > virsh console > No, not this. Putting correct lines in guest kernel command line I'm able to see what you wrote, directly from virt-manager through View --> Consoles --> Serial 0 I mean qemu monitor itself. The one that you activate inside qemu guest when you run qemu or qemu-kvm and then press Ctrl+ Alt + 2 and you get something like QEMU 0.11.0 monitor - type 'help' for more information (qemu) I understand that it would be risky to make things from the qemu monitor itself bypassing libvirt, and so I asked, just to know Thanks, Gianluca -------------- next part -------------- An HTML attachment was scrubbed... URL: From gianluca.cecchi at gmail.com Tue Oct 13 16:02:19 2009 From: gianluca.cecchi at gmail.com (Gianluca Cecchi) Date: Tue, 13 Oct 2009 18:02:19 +0200 Subject: [fedora-virt] duplicated console and details buttons Message-ID: <561c252c0910130902i19a76093h5b755b949458cf03@mail.gmail.com> You get the "console" and "details" buttons/functions both inside the toolbar and as labels just under the toolbar. It is ok that if you uncheck "view toolbar" you still get the labels, but it seems to me sort of redundant... they are right one below the same... see attached image Just my opinion. Gianluca -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: virt-manager.jpg Type: image/jpeg Size: 114226 bytes Desc: not available URL: From gianluca.cecchi at gmail.com Tue Oct 13 16:07:33 2009 From: gianluca.cecchi at gmail.com (Gianluca Cecchi) Date: Tue, 13 Oct 2009 18:07:33 +0200 Subject: [fedora-virt] status of guest not well visible when selected Message-ID: <561c252c0910130907y52ffbe71s5b33ead6c84f99@mail.gmail.com> So all is working like a charm and I have to scale to aesthetics.... ;-) With latest rawvirt repo on F11, when I have a guest selected in virt-manager gui, the current status of it (Running vs Shutoff) is pretty difficult to notice.... Ok, there is the play symbol inside the terminal icon but probably a black color for the state of the guest, at least when selected, would be more readable. Planning to scratch these two hosts of mine from F11 + rawvirt to F12 Beta... any insights/suggestions? Preupgrade option viable in your opinion? Bye and thanks, Gianluca -------------- next part -------------- An HTML attachment was scrubbed... URL: From crobinso at redhat.com Tue Oct 13 18:07:36 2009 From: crobinso at redhat.com (Cole Robinson) Date: Tue, 13 Oct 2009 14:07:36 -0400 Subject: [fedora-virt] duplicated console and details buttons In-Reply-To: <561c252c0910130902i19a76093h5b755b949458cf03@mail.gmail.com> References: <561c252c0910130902i19a76093h5b755b949458cf03@mail.gmail.com> Message-ID: <4AD4C1E8.50407@redhat.com> On 10/13/2009 12:02 PM, Gianluca Cecchi wrote: > You get the "console" and "details" buttons/functions both inside the > toolbar and as labels just under the toolbar. > It is ok that if you uncheck "view toolbar" you still get the labels, but it > seems to me sort of redundant... > they are right one below the same... see attached image > Hmm, those tabs should be hidden (they are for me on rawhide). What virt-manager and fedora version is this? - Cole From crobinso at redhat.com Tue Oct 13 18:09:04 2009 From: crobinso at redhat.com (Cole Robinson) Date: Tue, 13 Oct 2009 14:09:04 -0400 Subject: [fedora-virt] status of guest not well visible when selected In-Reply-To: <561c252c0910130907y52ffbe71s5b33ead6c84f99@mail.gmail.com> References: <561c252c0910130907y52ffbe71s5b33ead6c84f99@mail.gmail.com> Message-ID: <4AD4C240.7000901@redhat.com> On 10/13/2009 12:07 PM, Gianluca Cecchi wrote: > So all is working like a charm and I have to scale to aesthetics.... ;-) > > With latest rawvirt repo on F11, when I have a guest selected in > virt-manager gui, the current status of it (Running vs Shutoff) is pretty > difficult to notice.... Yes, I need to adjust this so that the text color changes to something more visible when the row is selected. - Cole From gianluca.cecchi at gmail.com Wed Oct 14 07:44:48 2009 From: gianluca.cecchi at gmail.com (Gianluca Cecchi) Date: Wed, 14 Oct 2009 09:44:48 +0200 Subject: [fedora-virt] duplicated console and details buttons In-Reply-To: <4AD4C1E8.50407@redhat.com> References: <561c252c0910130902i19a76093h5b755b949458cf03@mail.gmail.com> <4AD4C1E8.50407@redhat.com> Message-ID: <561c252c0910140044tb58325cv31a89d4c57fd550f@mail.gmail.com> On Tue, Oct 13, 2009 at 8:07 PM, Cole Robinson wrote: > > Hmm, those tabs should be hidden (they are for me on rawhide). What > virt-manager and fedora version is this? > > - Cole > > I'm on F11 x86_64 updated+ fedora-virt-preview repo. virt-manager-0.8.0-7.fc11.noarch libvirt-0.7.1-11.fc11.x86_64 qemu-0.11.0-6.fc11.x86_64 kernel-2.6.30.8-64.fc11.x86_64 Gianluca -------------- next part -------------- An HTML attachment was scrubbed... URL: From gianluca.cecchi at gmail.com Thu Oct 15 13:36:02 2009 From: gianluca.cecchi at gmail.com (Gianluca Cecchi) Date: Thu, 15 Oct 2009 15:36:02 +0200 Subject: [fedora-virt] USB passthrough and disconnect Message-ID: <561c252c0910150636p13705ca9p3caa0eb620ecb22f@mail.gmail.com> Hello, following the instructions at http://www.redhat.com/archives/fedora-virt/2009-June/msg00182.html I'm able to attach an USB key to my Windows XP SP3 guest. More better, I'm able to attach and use it with the guest already started. Very good! (Actually I don't have SElinux enabled so I skipped what related to it ) The doubt is, when in Windows I then safely remove the key and get the ok msg from the os, if I try to hot remove the USB device in details windows of the guest I receive the message that will be effective after reboot. Is this expected and the hot unplug a feature next to come, or do I miss anything in correct way to do things? My sw versions for host are F11 x86_64 with up2date rawvirt repo: libvirt-0.7.1-11.fc11.x86_64 qemu-kvm-0.11.0-6.fc11.x86_64 virt-manager-0.8.0-7.fc11.noarch Suppose the message is correct behaviour based on current state of things, what would be correct approach if during a vm session I use several usb keys for example? 1) insert first usb key in host 2) add hw of first usb key in details tab of guest 3) use it in guest and at the end disconnect it from inside guest OS 4) remove hw in details tab of guest; confirming that this will overwrite any othe rchanges occured in guest config.... 5) remove key form host usb plug 6) insert second usb key in host 7) add hw of second usb key in details tab of guest 8) use it in guest and at the end disconnect it from inside guest OS 9) ? If I do same as 4 do I loose the action made in 4) ?? Do I have to remove more hw components together only at the end? Any removal will let forget the previous ones? It seems not clear to me this message about " This device could not be removed from the running machine. Would you like to remove the device after the next VM shutdown? Warning: this will overwrite any other changes that require a VM reboot. " This is true for more than one usb key, but also for more than one device in general.... Thanks in advance, Gianluca -------------- next part -------------- An HTML attachment was scrubbed... URL: From crobinso at redhat.com Thu Oct 15 14:19:26 2009 From: crobinso at redhat.com (Cole Robinson) Date: Thu, 15 Oct 2009 10:19:26 -0400 Subject: [fedora-virt] USB passthrough and disconnect In-Reply-To: <561c252c0910150636p13705ca9p3caa0eb620ecb22f@mail.gmail.com> References: <561c252c0910150636p13705ca9p3caa0eb620ecb22f@mail.gmail.com> Message-ID: <4AD72F6E.3090405@redhat.com> On 10/15/2009 09:36 AM, Gianluca Cecchi wrote: > Hello, > following the instructions at > http://www.redhat.com/archives/fedora-virt/2009-June/msg00182.html > > I'm able to attach an USB key to my Windows XP SP3 guest. > More better, I'm able to attach and use it with the guest already started. > Very good! > (Actually I don't have SElinux enabled so I skipped what related to it ) > > The doubt is, when in Windows I then safely remove the key and get the ok > msg from the os, if I try to hot remove the USB device in details windows of > the guest I receive the message that will be effective after reboot. > Is this expected and the hot unplug a feature next to come, or do I miss > anything in correct way to do things? > USB hotunplug isn't supported by libvirt at the moment. I believe this is because it has historically been difficult to get the proper information from qemu that is required for hotunplug. > My sw versions for host are F11 x86_64 with up2date rawvirt repo: > libvirt-0.7.1-11.fc11.x86_64 > qemu-kvm-0.11.0-6.fc11.x86_64 > virt-manager-0.8.0-7.fc11.noarch > > Suppose the message is correct behaviour based on current state of things, > what would be correct approach if during a vm session I use several usb keys > for example? > 1) insert first usb key in host > 2) add hw of first usb key in details tab of guest > 3) use it in guest and at the end disconnect it from inside guest OS > 4) remove hw in details tab of guest; confirming that this will overwrite > any othe rchanges occured in guest config.... > 5) remove key form host usb plug > 6) insert second usb key in host > 7) add hw of second usb key in details tab of guest > 8) use it in guest and at the end disconnect it from inside guest OS > 9) ? If I do same as 4 do I loose the action made in 4) ?? > Yes, you will likely lose the effect of step 4. When the VM shuts down, USB1 will still appear to be attached. This is a side effect of how virt-manager interacts with libvirt to update a running VMs config when hotplug/unplug is not available (as in the USB hotunplug case, or say trying to hotplug an IDE disk). We actually should be able to fix in this in virt-manager now, and can remove the 'This will overwrite any other changes' text. I'll get to it for the next virt-manager release. - Cole From dlbewley at lib.ucdavis.edu Thu Oct 15 19:50:25 2009 From: dlbewley at lib.ucdavis.edu (Dale Bewley) Date: Thu, 15 Oct 2009 12:50:25 -0700 Subject: [fedora-virt] virtio block driver performance versus ide Message-ID: <1255636225.29670.40.camel@tofu.lib.ucdavis.edu> I assumed the virtio block driver to be faster for a kvm guest than ide, but I'm seeing quite the opposite. Is this expected? I normally create guests with a wrapper script around virt-install which creates guests configured to use IDE disks: f11-db1.xml- /usr/bin/qemu-kvm f11-db1.xml: f11-db1.xml- f11-db1.xml- f11-db1.xml: I am currently creating a guest with the virt-manager wizard which apparently uses virtio by default: f11-archivist.xml- /usr/bin/qemu-kvm f11-archivist.xml: f11-archivist.xml- f11-archivist.xml- f11-archivist.xml: The filesystem formatting and installation are going very very slowly. Host is 2.6.29.6-213.fc11.x86_64 4x quad core. Guests are also fc11 x86_64. BTW, I've updated to virt-manager-0.7.0-7, libvirt-0.6.2-18, qemu-kvm-0.10.6-6, and kernel-2.6.30.8-64 just before creating this guest but have not restarted libvirtd or rebooted the host. -- Dale Bewley - Unix Administrator - Shields Library - UC Davis GPG: 0xB098A0F3 0D5A 9AEB 43F4 F84C 7EFD 1753 064D 2583 B098 A0F3 From eb30750 at gmail.com Fri Oct 16 00:45:52 2009 From: eb30750 at gmail.com (Paul Lambert) Date: Thu, 15 Oct 2009 20:45:52 -0400 Subject: [fedora-virt] virtio block driver performance versus ide In-Reply-To: <1255636225.29670.40.camel@tofu.lib.ucdavis.edu> References: <1255636225.29670.40.camel@tofu.lib.ucdavis.edu> Message-ID: I will second this performance observation. After the latest round of pathces, I too found that my Windows XP VM was performing much faster than before and consuming much less cpu time. I then observed that my Fe-11 VM was performing considerably slower. Sure enough, I have checked my hda for each VM and the Windows was using native IDE and FE-11 was using virtio. I have now changed my FE-11 hda to IDE. Seems to exhibit much more pep! So much for paravirtualization. Paul On Thu, Oct 15, 2009 at 3:50 PM, Dale Bewley wrote: > I assumed the virtio block driver to be faster for a kvm guest than ide, > but I'm seeing quite the opposite. Is this expected? > > I normally create guests with a wrapper script around virt-install which > creates guests configured to use IDE disks: > > f11-db1.xml- /usr/bin/qemu-kvm > f11-db1.xml: > f11-db1.xml- > f11-db1.xml- > f11-db1.xml: > > I am currently creating a guest with the virt-manager wizard which > apparently uses virtio by default: > > f11-archivist.xml- /usr/bin/qemu-kvm > f11-archivist.xml: > f11-archivist.xml- file='/var/lib/libvirt/images//f11-archivist.img'/> > f11-archivist.xml- > f11-archivist.xml: > > The filesystem formatting and installation are going very very slowly. > > Host is 2.6.29.6-213.fc11.x86_64 4x quad core. Guests are also fc11 > x86_64. > > BTW, I've updated to virt-manager-0.7.0-7, libvirt-0.6.2-18, > qemu-kvm-0.10.6-6, and kernel-2.6.30.8-64 just before creating this > guest but have not restarted libvirtd or rebooted the host. > > -- > Dale Bewley - Unix Administrator - Shields Library - UC Davis > GPG: 0xB098A0F3 0D5A 9AEB 43F4 F84C 7EFD 1753 064D 2583 B098 A0F3 > > _______________________________________________ > Fedora-virt mailing list > Fedora-virt at redhat.com > https://www.redhat.com/mailman/listinfo/fedora-virt > -------------- next part -------------- An HTML attachment was scrubbed... URL: From rjones at redhat.com Fri Oct 16 10:04:58 2009 From: rjones at redhat.com (Richard W.M. Jones) Date: Fri, 16 Oct 2009 11:04:58 +0100 Subject: [fedora-virt] Re: nagios-virt: configure: error: Cannot find `nagios' program In-Reply-To: <5403369.1255685978676.JavaMail.www@wwumf0102> References: <5403369.1255685978676.JavaMail.www@wwumf0102> Message-ID: <20091016100458.GN24800@amd.home.annexia.org> On Fri, Oct 16, 2009 at 11:39:38AM +0200, Stephane Grand wrote: > Hello, > > I've found your web page about nagios-virt (http://et.redhat.com/~rjones/nagios-virt/). This tools interesting me to supervise servers and vm. > > Unfortunately, I can't test it. Here is error message I got when I execute "./configure" > > [root] # ./configure > ... > checking for stdint.h... yes > checking for unistd.h... yes > checking libvirt/libvirt.h usability... yes > checking libvirt/libvirt.h presence... yes > checking for libvirt/libvirt.h... yes > checking for virConnectGetCapabilities... yes > checking for nagios... no > configure: error: Cannot find `nagios' program > > I've retried this command with others options but I got the same error message Does the 'nagios' command exist? Is it on the $PATH? Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones libguestfs lets you edit virtual machines. Supports shell scripting, bindings from many languages. http://et.redhat.com/~rjones/libguestfs/ See what it can do: http://et.redhat.com/~rjones/libguestfs/recipes.html From james.mcevoy at hp.com Fri Oct 16 15:21:06 2009 From: james.mcevoy at hp.com (McEvoy, James) Date: Fri, 16 Oct 2009 15:21:06 +0000 Subject: [fedora-virt] Re: nagios-virt: configure: error: Cannot find `nagios' program In-Reply-To: <20091016100458.GN24800@amd.home.annexia.org> References: <5403369.1255685978676.JavaMail.www@wwumf0102> <20091016100458.GN24800@amd.home.annexia.org> Message-ID: <197A4D1DF629F64A9260E3B454A218D134180913A9@GVW1158EXB.americas.hpqcorp.net> There is a nagios package for system monitoring on Fedora 11 in the updates repository. Available Packages Name : nagios Arch : x86_64 Version : 3.2.0 Release : 2.fc11 Size : 3.6 M Repo : updates Summary : Nagios monitors hosts and services and yells if somethings breaks URL : http://www.nagios.org/ License : GPLv2 Description: Nagios is a program that will monitor hosts and services on your : network. It has the ability to send email or page alerts when a : problem arises and when a problem is resolved. Nagios is written : in C and is designed to run under Linux (and some other *NIX : variants) as a background process, intermittently running checks : on various services that you specify. : : The actual service checks are performed by separate "plugin" : programs which return the status of the checks to Nagios. The : plugins are available at : http://sourceforge.net/projects/nagiosplug. : : This package provides the core program, web interface, and : documentation files for Nagios. Development files are built as a : separate package. -----Original Message----- From: fedora-virt-bounces at redhat.com [mailto:fedora-virt-bounces at redhat.com] On Behalf Of Richard W.M. Jones Sent: Friday, October 16, 2009 3:05 AM To: Stephane Grand; fedora-virt at redhat.com Subject: [fedora-virt] Re: nagios-virt: configure: error: Cannot find `nagios' program On Fri, Oct 16, 2009 at 11:39:38AM +0200, Stephane Grand wrote: > Hello, > > I've found your web page about nagios-virt (http://et.redhat.com/~rjones/nagios-virt/). This tools interesting me to supervise servers and vm. > > Unfortunately, I can't test it. Here is error message I got when I execute "./configure" > > [root] # ./configure > ... > checking for stdint.h... yes > checking for unistd.h... yes > checking libvirt/libvirt.h usability... yes > checking libvirt/libvirt.h presence... yes > checking for libvirt/libvirt.h... yes > checking for virConnectGetCapabilities... yes > checking for nagios... no > configure: error: Cannot find `nagios' program > > I've retried this command with others options but I got the same error message Does the 'nagios' command exist? Is it on the $PATH? Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones libguestfs lets you edit virtual machines. Supports shell scripting, bindings from many languages. http://et.redhat.com/~rjones/libguestfs/ See what it can do: http://et.redhat.com/~rjones/libguestfs/recipes.html _______________________________________________ Fedora-virt mailing list Fedora-virt at redhat.com https://www.redhat.com/mailman/listinfo/fedora-virt From dlbewley at lib.ucdavis.edu Fri Oct 16 21:11:37 2009 From: dlbewley at lib.ucdavis.edu (Dale Bewley) Date: Fri, 16 Oct 2009 14:11:37 -0700 (PDT) Subject: [fedora-virt] virtio block driver performance versus ide In-Reply-To: Message-ID: <957322961.1361301255727497547.JavaMail.root@zebra.lib.ucdavis.edu> I've installed two F11 guests of the same config other than IDE vs VirtIO disks. I updated them fully, and they are performing at comparable speed. Perhaps the issue is only with the kernel in the installer image. -- Dale Bewley - Unix Administrator - Shields Library - UC Davis GPG: 0xB098A0F3 0D5A 9AEB 43F4 F84C 7EFD 1753 064D 2583 B098 A0F3 From mbooth at redhat.com Mon Oct 19 09:47:13 2009 From: mbooth at redhat.com (Matthew Booth) Date: Mon, 19 Oct 2009 10:47:13 +0100 Subject: [fedora-virt] 'eject' on a KVM guest Message-ID: <4ADC35A1.1060005@redhat.com> What does the eject (to eject a CD) command do to a KVM guest? It succeeds, and the result is that the device can no longer be mounted. However, looking at the domain XML reveals it hasn't disconnected the backing file. Matt -- Matthew Booth, RHCA, RHCSS Red Hat Engineering, Virtualisation Team M: +44 (0)7977 267231 GPG ID: D33C3490 GPG FPR: 3733 612D 2D05 5458 8A8A 1600 3441 EA19 D33C 3490 From pavel.lisy at gmail.com Mon Oct 19 10:37:32 2009 From: pavel.lisy at gmail.com (Pavel Lisy) Date: Mon, 19 Oct 2009 12:37:32 +0200 Subject: [fedora-virt] how to edit dhcp configuration Message-ID: <1255948652.7408.46.camel@pali-pc.hk.tmapy.cz> Hello I want to set fixed IP for host (from MAC address) in dhcp. How can I do it through libvirt configuration? There is not possibility set it in virt-manager. I can edit manually ldapnet.xml (ldapnet is name of my network) but: 1. question: Where is right place to edit it: /etc/libvirt/qemu/networks/ldapnet.xml or /var/lib/libvirt/network/ldapnet.xml (after reboot one file replaces other but which direction but I didn't find which direction) 2. question: How can I propagate this changes? Neither /etc/init.d/libvirtd reload /etc/init.d/libvirtd restart changes dnsmasq settings. It is done only during reboot, why? 3. I've found this workaround: * kill dnsmasq process for this settings * start dnsmasq with new settings, ex. .... --dhcp-host 52:54:00:22:a7:76,ldap-m1.ldapnet.tmapy.cz,192.168.241.101 Is it way how to do it through libvirt utility? Pavel -- Pavel Lisy From markmc at redhat.com Mon Oct 19 11:31:50 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Mon, 19 Oct 2009 12:31:50 +0100 Subject: [fedora-virt] virt-preview updates In-Reply-To: <1250703309.17874.46.camel@blaa> References: <1244034684.5001.134.camel@blaa> <1244046483.5001.180.camel@blaa> <1244197459.27876.0.camel@blaa> <1246303956.11688.148.camel@blaa> <1246628572.13604.2.camel@blaa> <1246636154.13604.8.camel@blaa> <1247763853.3038.105.camel@blaa> <1248769817.3089.20.camel@blaa> <1248776732.3089.45.camel@blaa> <1248859649.3021.28.camel@blaa> <20090729093637.GC10723@salstar.sk> <1248868438.3021.31.camel@blaa> <1248873640.3021.34.camel@blaa> <1248879462.3021.39.camel@blaa> <1248882336.3021.40.camel@blaa> <1248974623.3275.4.camel@blaa> <1249057022.3320.35.camel@blaa> <1249400197.3212.73.camel@blaa> <1249573056.3721.62.camel@blaa> <1249639097.3260.5.camel@blaa> <1249900612.8784.79.camel@blaa> <1250700893.17874.44.camel@blaa> <1250703309.17874.46.camel@blaa> Message-ID: <1255951910.18316.12.camel@blaa> The latest updates available from http://markmc.fedorapeople.org/virt-preview == qemu == * Mon Oct 19 2009 Mark McLoughlin - 2:0.11.0-7 - Fix potential segfault from too small MSR_COUNT (#528901) == libvirt == * Mon Oct 19 2009 Mark McLoughlin - 0.7.1-13 - Misc fixes to qemu machine types handling - A couple of XML formatting fixes * Tue Oct 13 2009 Mark McLoughlin - 0.7.1-12 - Fix restore of qemu guest using raw save format (#523158) Cheers, Mark. == About The Virtualization Preview Repository == The virt-preview repository offers Fedora 11 users the opportunity to help out with testing the rawhide virtualization packages, without having to run all of rawhide. Bugs should be reported to http://bugzilla.redhat.com/ using 'rawhide' as the distribution version. Note clearly in the bug report that you are running these packages on Fedora 11. From crobinso at redhat.com Mon Oct 19 13:35:04 2009 From: crobinso at redhat.com (Cole Robinson) Date: Mon, 19 Oct 2009 09:35:04 -0400 Subject: [fedora-virt] 'eject' on a KVM guest In-Reply-To: <4ADC35A1.1060005@redhat.com> References: <4ADC35A1.1060005@redhat.com> Message-ID: <4ADC6B08.1020504@redhat.com> On 10/19/2009 05:47 AM, Matthew Booth wrote: > What does the eject (to eject a CD) command do to a KVM guest? It > succeeds, and the result is that the device can no longer be mounted. > However, looking at the domain XML reveals it hasn't disconnected the > backing file. > > Matt It should be the equivalent of ejecting a CD or floppy on a physical machine. If you do this through libvirt (with the attach-device command and suitable XML string), then the change should be reflected in the XML. If you managed to send a monitor command to the KVM guest that didn't go through libvirt, the change wouldn't be reflected in the domain XML for the majority of operations (media eject included). - Cole From berrange at redhat.com Mon Oct 19 13:38:57 2009 From: berrange at redhat.com (Daniel P. Berrange) Date: Mon, 19 Oct 2009 14:38:57 +0100 Subject: [fedora-virt] 'eject' on a KVM guest In-Reply-To: <4ADC6B08.1020504@redhat.com> References: <4ADC35A1.1060005@redhat.com> <4ADC6B08.1020504@redhat.com> Message-ID: <20091019133857.GK27871@redhat.com> On Mon, Oct 19, 2009 at 09:35:04AM -0400, Cole Robinson wrote: > On 10/19/2009 05:47 AM, Matthew Booth wrote: > > What does the eject (to eject a CD) command do to a KVM guest? It > > succeeds, and the result is that the device can no longer be mounted. > > However, looking at the domain XML reveals it hasn't disconnected the > > backing file. > > > > Matt > > It should be the equivalent of ejecting a CD or floppy on a physical > machine. > > If you do this through libvirt (with the attach-device command and > suitable XML string), then the change should be reflected in the XML. > > If you managed to send a monitor command to the KVM guest that didn't go > through libvirt, the change wouldn't be reflected in the domain XML for > the majority of operations (media eject included). I think the problem Matt sees is from running 'eject' inside the guest. libvirt will not see that, because KVM has no way to notify us that the guest ejected the media. Fortunately, the new KVM monitor control protocol includes support for this notification, so we'll be able to address this limitation in the future. Daniel -- |: Red Hat, Engineering, London -o- http://people.redhat.com/berrange/ :| |: http://libvirt.org -o- http://virt-manager.org -o- http://ovirt.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: GnuPG: 7D3B9505 -o- F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 :| From crobinso at redhat.com Mon Oct 19 13:39:44 2009 From: crobinso at redhat.com (Cole Robinson) Date: Mon, 19 Oct 2009 09:39:44 -0400 Subject: [fedora-virt] how to edit dhcp configuration In-Reply-To: <1255948652.7408.46.camel@pali-pc.hk.tmapy.cz> References: <1255948652.7408.46.camel@pali-pc.hk.tmapy.cz> Message-ID: <4ADC6C20.30907@redhat.com> On 10/19/2009 06:37 AM, Pavel Lisy wrote: > Hello > > I want to set fixed IP for host (from MAC address) in dhcp. How can I do > it through libvirt configuration? > > There is not possibility set it in virt-manager. I can edit manually > ldapnet.xml (ldapnet is name of my network) but: > > 1. question: > Where is right place to edit it: > /etc/libvirt/qemu/networks/ldapnet.xml > or > /var/lib/libvirt/network/ldapnet.xml > (after reboot one file replaces other but which direction but I didn't > find which direction) > The 'right' place is to use 'virsh net-edit '. Editing the files in place provides no error checking and can require a libvirtd restart. The user shouldn't have to know where the files are kept. > 2. question: > How can I propagate this changes? > > Neither > /etc/init.d/libvirtd reload > /etc/init.d/libvirtd restart > changes dnsmasq settings. It is done only during reboot, why? > Before you make the changes, do 'virsh net-destroy '. This will shutdown the network (dnsmasq), but NOT delete it in any way. Then use 'virsh net-edit ' as mentioned above to make your changes. Finally, use 'virsh net-start ' which will start a new dnsmasq process. - Cole > 3. I've found this workaround: > * kill dnsmasq process for this settings > * start dnsmasq with new settings, ex. > .... --dhcp-host 52:54:00:22:a7:76,ldap-m1.ldapnet.tmapy.cz,192.168.241.101 > Is it way how to do it through libvirt utility? > > Pavel From mbooth at redhat.com Mon Oct 19 13:49:52 2009 From: mbooth at redhat.com (Matthew Booth) Date: Mon, 19 Oct 2009 14:49:52 +0100 Subject: [fedora-virt] 'eject' on a KVM guest In-Reply-To: <20091019133857.GK27871@redhat.com> References: <4ADC35A1.1060005@redhat.com> <4ADC6B08.1020504@redhat.com> <20091019133857.GK27871@redhat.com> Message-ID: <4ADC6E80.1090606@redhat.com> On 19/10/09 14:38, Daniel P. Berrange wrote: > On Mon, Oct 19, 2009 at 09:35:04AM -0400, Cole Robinson wrote: >> On 10/19/2009 05:47 AM, Matthew Booth wrote: >>> What does the eject (to eject a CD) command do to a KVM guest? It >>> succeeds, and the result is that the device can no longer be mounted. >>> However, looking at the domain XML reveals it hasn't disconnected the >>> backing file. >>> >>> Matt >> >> It should be the equivalent of ejecting a CD or floppy on a physical >> machine. >> >> If you do this through libvirt (with the attach-device command and >> suitable XML string), then the change should be reflected in the XML. >> >> If you managed to send a monitor command to the KVM guest that didn't go >> through libvirt, the change wouldn't be reflected in the domain XML for >> the majority of operations (media eject included). > > I think the problem Matt sees is from running 'eject' inside the guest. > libvirt will not see that, because KVM has no way to notify us that the > guest ejected the media. That's it. Re-reading, that wasn't clear from my original question. > Fortunately, the new KVM monitor control protocol includes support for > this notification, so we'll be able to address this limitation in the > future. Sounds good. Any idea what the timeframe might be for that? To put this in context, I was trying to think of an existing way a guest can modify its environment. I was thinking that if you configured a boot order of cdrom, hd, and attached a custom LiveCD, the LiveCD could do its thing and then eject itself, booting from hd next time with no administrative intervention. Could that ever work? Matt -- Matthew Booth, RHCA, RHCSS Red Hat Engineering, Virtualisation Team M: +44 (0)7977 267231 GPG ID: D33C3490 GPG FPR: 3733 612D 2D05 5458 8A8A 1600 3441 EA19 D33C 3490 From berrange at redhat.com Mon Oct 19 13:51:38 2009 From: berrange at redhat.com (Daniel P. Berrange) Date: Mon, 19 Oct 2009 14:51:38 +0100 Subject: [fedora-virt] 'eject' on a KVM guest In-Reply-To: <4ADC6E80.1090606@redhat.com> References: <4ADC35A1.1060005@redhat.com> <4ADC6B08.1020504@redhat.com> <20091019133857.GK27871@redhat.com> <4ADC6E80.1090606@redhat.com> Message-ID: <20091019135138.GL27871@redhat.com> On Mon, Oct 19, 2009 at 02:49:52PM +0100, Matthew Booth wrote: > On 19/10/09 14:38, Daniel P. Berrange wrote: > >On Mon, Oct 19, 2009 at 09:35:04AM -0400, Cole Robinson wrote: > >>On 10/19/2009 05:47 AM, Matthew Booth wrote: > >>>What does the eject (to eject a CD) command do to a KVM guest? It > >>>succeeds, and the result is that the device can no longer be mounted. > >>>However, looking at the domain XML reveals it hasn't disconnected the > >>>backing file. > >>> > >>>Matt > >> > >>It should be the equivalent of ejecting a CD or floppy on a physical > >>machine. > >> > >>If you do this through libvirt (with the attach-device command and > >>suitable XML string), then the change should be reflected in the XML. > >> > >>If you managed to send a monitor command to the KVM guest that didn't go > >>through libvirt, the change wouldn't be reflected in the domain XML for > >>the majority of operations (media eject included). > > > >I think the problem Matt sees is from running 'eject' inside the guest. > >libvirt will not see that, because KVM has no way to notify us that the > >guest ejected the media. > > That's it. Re-reading, that wasn't clear from my original question. > > >Fortunately, the new KVM monitor control protocol includes support for > >this notification, so we'll be able to address this limitation in the > >future. > > Sounds good. Any idea what the timeframe might be for that? When QEMU developers agree on the implementation of the new protocol... > > To put this in context, I was trying to think of an existing way a guest > can modify its environment. I was thinking that if you configured a boot > order of cdrom, hd, and attached a custom LiveCD, the LiveCD could do > its thing and then eject itself, booting from hd next time with no > administrative intervention. Could that ever work? Yes, that kind of thing is supposed to work. Daniel -- |: Red Hat, Engineering, London -o- http://people.redhat.com/berrange/ :| |: http://libvirt.org -o- http://virt-manager.org -o- http://ovirt.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: GnuPG: 7D3B9505 -o- F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 :| From stephane.grand at atosorigin.com Fri Oct 16 14:09:23 2009 From: stephane.grand at atosorigin.com (Stephane Grand) Date: Fri, 16 Oct 2009 16:09:23 +0200 (CEST) Subject: [fedora-virt] Re: Re: nagios-virt: configure: error: Cannot find `nagios' program Message-ID: <27926988.1255702163603.JavaMail.www@wwumf0101> Thank you for the answer. Here is what I have done to follow the installation process. [root] # PATH=$PATH:/var/supervision/nagios/bin/ [root] # ./configure --with-nagios-configdir=/var/supervision/nagios/etc [root] # make [root] # ./nagios-virt config PACKAGE_NAME nagios-virt PACKAGE_VERSION 0.2.0 NAGIOS_CONFIGDIR /var/supervision/nagios/etc/ NV_DATADIR /usr/local/share/nagios-virt-0.2.0 ==>But I've got another problem. [root] # ./nagios-virt install Installing virt-library.cfg in /var/supervision/nagios/etc/ ... Installing /var/supervision/nagios/etc//commands.cfg ... Warning: could not install /var/supervision/nagios/etc//commands.cfg This file from the standard Nagios distribution is required in order for nagios-virt to work correctly. Installing virt-hosts.cfg in /var/supervision/nagios/etc/ ... virConnectNumOfDomains: failed Why this command can not install "commands.cfg" file ? Thank you. Stephane ======================================== Date du message : oct. 16 2009, 12:05 PM De : "Richard W.M. Jones" A : "Stephane Grand" , fedora-virt at redhat.com Copie : Sujet : Re: nagios-virt: configure: error: Cannot find `nagios' program On Fri, Oct 16, 2009 at 11:39:38AM +0200, Stephane Grand wrote: > Hello, > > I've found your web page about nagios-virt (http://et.redhat.com/~rjones/nagios-virt/). This tools interesting me to supervise servers and vm. > > Unfortunately, I can't test it. Here is error message I got when I execute "./configure" > > [root] # ./configure > ... > checking for stdint.h... yes > checking for unistd.h... yes > checking libvirt/libvirt.h usability... yes > checking libvirt/libvirt.h presence... yes > checking for libvirt/libvirt.h... yes > checking for virConnectGetCapabilities... yes > checking for nagios... no > configure: error: Cannot find `nagios' program > > I've retried this command with others options but I got the same error message Does the 'nagios' command exist? Is it on the $PATH? Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones libguestfs lets you edit virtual machines. Supports shell scripting, bindings from many languages. http://et.redhat.com/~rjones/libguestfs/ See what it can do: http://et.redhat.com/~rjones/libguestfs/recipes.html ----------------------- Atos Origin St?phane Grand Open Source Center dcen-net-grand.consultant at dgfip.finances.gouv.fr From gvarisco at redhat.com Mon Oct 19 18:15:51 2009 From: gvarisco at redhat.com (Gianluca Varisco) Date: Mon, 19 Oct 2009 20:15:51 +0200 Subject: [fedora-virt] #529711 - qemu: could not open serial device 'pty' Message-ID: <4ADCACD7.3010407@redhat.com> Hi guys, Am I the only one having this issue with qemu/KVM and rawhide? https://bugzilla.redhat.com/show_bug.cgi?id=529711 (It is not a SELinux-related problem) Cheers, -- Gianluca Varisco, RHCE | Office: +39 02 9737 4652 Red Hat Italia | Fax: +39 02 669 3111 Via Antonio Da Recanate 1 | Mobile: +39 333 574 0934 20124 Milano | eMail: gvarisco at redhat.com From pasik at iki.fi Mon Oct 19 18:39:25 2009 From: pasik at iki.fi (Pasi =?iso-8859-1?Q?K=E4rkk=E4inen?=) Date: Mon, 19 Oct 2009 21:39:25 +0300 Subject: [fedora-virt] virtio block driver performance versus ide In-Reply-To: References: <1255636225.29670.40.camel@tofu.lib.ucdavis.edu> Message-ID: <20091019183925.GO1434@reaktio.net> On Thu, Oct 15, 2009 at 08:45:52PM -0400, Paul Lambert wrote: > I will second this performance observation. After the latest round of > pathces, I too found that my Windows XP VM was performing much faster than > before and consuming much less cpu time. I then observed that my Fe-11 VM > was performing considerably slower. Sure enough, I have checked my hda > for each VM and the Windows was using native IDE and FE-11 was using > virtio. I have now changed my FE-11 hda to IDE. Seems to exhibit much > more pep! So much for paravirtualization. > It sounds like the kvm virtio drivers are not working like they should. possible a bug in the drivers.. At least Xen PV drivers are faster than emulated IDE, and use less CPU. -- Pasi > Paul > > On Thu, Oct 15, 2009 at 3:50 PM, Dale Bewley <[1]dlbewley at lib.ucdavis.edu> > wrote: > > I assumed the virtio block driver to be faster for a kvm guest than ide, > but I'm seeing quite the opposite. Is this expected? > > I normally create guests with a wrapper script around virt-install which > creates guests configured to use IDE disks: > > f11-db1.xml- /usr/bin/qemu-kvm > f11-db1.xml: > f11-db1.xml- > f11-db1.xml- > f11-db1.xml: > > I am currently creating a guest with the virt-manager wizard which > apparently uses virtio by default: > > f11-archivist.xml- /usr/bin/qemu-kvm > f11-archivist.xml: > f11-archivist.xml- file='/var/lib/libvirt/images//f11-archivist.img'/> > f11-archivist.xml- > f11-archivist.xml: > > The filesystem formatting and installation are going very very slowly. > > Host is 2.6.29.6-213.fc11.x86_64 4x quad core. Guests are also fc11 > x86_64. > > BTW, I've updated to virt-manager-0.7.0-7, libvirt-0.6.2-18, > qemu-kvm-0.10.6-6, and kernel-2.6.30.8-64 just before creating this > guest but have not restarted libvirtd or rebooted the host. > -- > Dale Bewley - Unix Administrator - Shields Library - UC Davis > GPG: 0xB098A0F3 0D5A 9AEB 43F4 F84C 7EFD 1753 064D 2583 B098 A0F3 > > _______________________________________________ > Fedora-virt mailing list > [2]Fedora-virt at redhat.com > [3]https://www.redhat.com/mailman/listinfo/fedora-virt > > References > > Visible links > 1. mailto:dlbewley at lib.ucdavis.edu > 2. mailto:Fedora-virt at redhat.com > 3. https://www.redhat.com/mailman/listinfo/fedora-virt > _______________________________________________ > Fedora-virt mailing list > Fedora-virt at redhat.com > https://www.redhat.com/mailman/listinfo/fedora-virt From dlaor at redhat.com Tue Oct 20 08:44:58 2009 From: dlaor at redhat.com (Dor Laor) Date: Tue, 20 Oct 2009 10:44:58 +0200 Subject: [fedora-virt] virtio block driver performance versus ide In-Reply-To: <20091019183925.GO1434@reaktio.net> References: <1255636225.29670.40.camel@tofu.lib.ucdavis.edu> <20091019183925.GO1434@reaktio.net> Message-ID: <4ADD788A.6090000@redhat.com> On 10/19/2009 08:39 PM, Pasi K?rkk?inen wrote: > On Thu, Oct 15, 2009 at 08:45:52PM -0400, Paul Lambert wrote: >> I will second this performance observation. After the latest round of >> pathces, I too found that my Windows XP VM was performing much faster than >> before and consuming much less cpu time. I then observed that my Fe-11 VM >> was performing considerably slower. Sure enough, I have checked my hda >> for each VM and the Windows was using native IDE and FE-11 was using >> virtio. I have now changed my FE-11 hda to IDE. Seems to exhibit much >> more pep! So much for paravirtualization. >> > > It sounds like the kvm virtio drivers are not working like they should. > possible a bug in the drivers.. > > At least Xen PV drivers are faster than emulated IDE, and use less CPU. WinXp is an exception for kvm virtio drivers. More info at http://www.mail-archive.com/kvm at vger.kernel.org/msg22834.html > > -- Pasi > >> Paul >> >> On Thu, Oct 15, 2009 at 3:50 PM, Dale Bewley<[1]dlbewley at lib.ucdavis.edu> >> wrote: >> >> I assumed the virtio block driver to be faster for a kvm guest than ide, >> but I'm seeing quite the opposite. Is this expected? >> >> I normally create guests with a wrapper script around virt-install which >> creates guests configured to use IDE disks: >> >> f11-db1.xml-/usr/bin/qemu-kvm >> f11-db1.xml: >> f11-db1.xml- >> f11-db1.xml- >> f11-db1.xml: >> >> I am currently creating a guest with the virt-manager wizard which >> apparently uses virtio by default: >> >> f11-archivist.xml-/usr/bin/qemu-kvm >> f11-archivist.xml: >> f11-archivist.xml-> file='/var/lib/libvirt/images//f11-archivist.img'/> >> f11-archivist.xml- >> f11-archivist.xml: >> >> The filesystem formatting and installation are going very very slowly. >> >> Host is 2.6.29.6-213.fc11.x86_64 4x quad core. Guests are also fc11 >> x86_64. >> >> BTW, I've updated to virt-manager-0.7.0-7, libvirt-0.6.2-18, >> qemu-kvm-0.10.6-6, and kernel-2.6.30.8-64 just before creating this >> guest but have not restarted libvirtd or rebooted the host. >> -- >> Dale Bewley - Unix Administrator - Shields Library - UC Davis >> GPG: 0xB098A0F3 0D5A 9AEB 43F4 F84C 7EFD 1753 064D 2583 B098 A0F3 >> >> _______________________________________________ >> Fedora-virt mailing list >> [2]Fedora-virt at redhat.com >> [3]https://www.redhat.com/mailman/listinfo/fedora-virt >> >> References >> >> Visible links >> 1. mailto:dlbewley at lib.ucdavis.edu >> 2. mailto:Fedora-virt at redhat.com >> 3. https://www.redhat.com/mailman/listinfo/fedora-virt > >> _______________________________________________ >> Fedora-virt mailing list >> Fedora-virt at redhat.com >> https://www.redhat.com/mailman/listinfo/fedora-virt > > _______________________________________________ > Fedora-virt mailing list > Fedora-virt at redhat.com > https://www.redhat.com/mailman/listinfo/fedora-virt From gianluca.cecchi at gmail.com Tue Oct 20 12:33:03 2009 From: gianluca.cecchi at gmail.com (Gianluca Cecchi) Date: Tue, 20 Oct 2009 14:33:03 +0200 Subject: [fedora-virt] #529711 - qemu: could not open serial device 'pty' Message-ID: <561c252c0910200533g7eb51d5bwe79f498c5384af86@mail.gmail.com> On Mon, 19 Oct 2009 20:15:51 +0200 Gianluca Varisco wrote: > Hi guys, > > Am I the only one having this issue with qemu/KVM and rawhide? > > https://bugzilla.redhat.com/show_bug.cgi?id=529711 > > (It is not a SELinux-related problem) > > Cheers, I would advise to also follow what written at link: https://fedoraproject.org/wiki/Reporting_virtualization_bugs In particular contents of virt-manager.log and posting output when setting LIBVIRT_DEBUG=1 btw I presume that also the version of virt-manager is the latest one, correct? HIH, Gianluca -------------- next part -------------- An HTML attachment was scrubbed... URL: From gianluca.cecchi at gmail.com Tue Oct 20 15:26:24 2009 From: gianluca.cecchi at gmail.com (Gianluca Cecchi) Date: Tue, 20 Oct 2009 17:26:24 +0200 Subject: [fedora-virt] virt-manager and ssh with non root user Message-ID: <561c252c0910200826l6d338f22tdd58360d7d3fec08@mail.gmail.com> Hello, using F11 + rawvirt repo. I would like to connect from my virt-manager from this F11 to a RH EL 5.4 xen Hypervisor. Is it possible? On the RH EL host, root access via ssh is prevented and I don't want to use a CA certficate at the moment. Also I cannot export DISPLAY env so I cannot directly ssh to the host and run its virt-manager program.... I tried to edit /etc/libvirtd/libvirtd.conf on rh el 5.4 host adding unix_sock_group = "mygroup" and then trying to connect with one user of this group. Is this setting correct as I intended it? When in virt-manager of F11 I try to create new connection, it seems I'm only able to create a session of type xen+ssh://root at myrhel54host/ and not of type xen+ssh://myuser at myrhel54host Any hints? Also, as in virt-manager this connection fails to create (root prevented from ssh directly as said above), I would like to delete it... but I didn't find an option for this; In virt-manager only "connect" and "details" options are available.... these are messages in virsh session when setting LIBVIRT_DEBUG=1 virsh # connect xen+ssh://unix at myrhel54host 17:23:44.911: debug : virConnectClose:1291 : conn=0x65de70 17:23:44.911: debug : virUnrefConnect:257 : unref connection 0x65de70 1 17:23:44.911: debug : remoteIO:7421 : Do proc=2 serial=3 length=28 wait=(nil) 17:23:44.911: debug : remoteIO:7483 : We have the buck 2 0x67db20 0x67db20 17:23:44.912: debug : remoteIODecodeMessageLength:7032 : Got length, now need 56 total (52 more) 17:23:44.912: debug : remoteIOEventLoop:7347 : Giving up the buck 2 0x67db20 (nil) 17:23:44.912: debug : remoteIO:7514 : All done with our call 2 (nil) 0x67db20 17:23:44.912: debug : virReleaseConnect:214 : release connection 0x65de70 17:23:44.912: debug : virConnectOpenAuth:1273 : name=xen+ssh://unix at myrhel54host, auth=0x7fb212bc40a0, flags=0 17:23:44.912: debug : do_open:1042 : name "xen+ssh://unix at myrhel54host" to URI components: scheme xen+ssh opaque (null) authority (null) server myrhel54host user unix port 0 path (null) 17:23:44.912: debug : do_open:1052 : trying driver 0 (Test) ... 17:23:44.912: debug : do_open:1058 : driver 0 Test returned DECLINED 17:23:44.912: debug : do_open:1052 : trying driver 1 (Xen) ... 17:23:44.912: debug : do_open:1058 : driver 1 Xen returned DECLINED 17:23:44.912: debug : do_open:1052 : trying driver 2 (OPENVZ) ... 17:23:44.912: debug : do_open:1058 : driver 2 OPENVZ returned DECLINED 17:23:44.912: debug : do_open:1052 : trying driver 3 (PHYP) ... 17:23:44.912: debug : do_open:1058 : driver 3 PHYP returned DECLINED 17:23:44.912: debug : do_open:1052 : trying driver 4 (VBOX) ... 17:23:44.912: debug : do_open:1058 : driver 4 VBOX returned DECLINED 17:23:44.912: debug : do_open:1052 : trying driver 5 (ESX) ... 17:23:44.912: debug : do_open:1058 : driver 5 ESX returned DECLINED 17:23:44.912: debug : do_open:1052 : trying driver 6 (remote) ... 17:23:44.912: debug : doRemoteOpen:535 : proceeding with name = xen:// 17:23:44.912: debug : virExecWithHook:609 : ssh -l unix myrhel54host nc -U /var/run/libvirt/libvirt-sock 17:23:44.917: debug : remoteIO:7421 : Do proc=66 serial=0 length=28 wait=(nil) 17:23:44.918: debug : remoteIO:7483 : We have the buck 66 0x6bde80 0x6bde80 17:23:44.924: debug : remoteIOEventLoop:7368 : Giving up the buck due to I/O error 66 0x6bde80 (nil) 17:23:44.924: debug : do_open:1058 : driver 6 remote returned ERROR 17:23:44.924: debug : virUnrefConnect:257 : unref connection 0x65e260 1 17:23:44.924: debug : virReleaseConnect:214 : release connection 0x65e260 error: Failed to connect to the hypervisor error: cannot recv data: Connection reset by peer Thanks, Gianluca -------------- next part -------------- An HTML attachment was scrubbed... URL: From markmc at redhat.com Wed Oct 21 07:20:49 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Wed, 21 Oct 2009 08:20:49 +0100 Subject: [fedora-virt] Re: Oops In-Reply-To: <200910201939.05147.gene@czarc.net> References: <200910201939.05147.gene@czarc.net> Message-ID: <1256109649.2861.4.camel@blaa> Hi Gene, On Tue, 2009-10-20 at 19:39 -0400, Gene Czarcinski wrote: > F12 beta was available so I downloaded the DVD iso ... that was a waste of > bandwidth since it was identical to F12-Beta RC2. > > Anyway, I had already installed a couple of copies of RC2 into qemu-kvm guests > so it was time to install this on bare-metal. > > My test system already had F12-rawhide installed with all current updates but, > afterall, this is a test of the install ... so I did another install ... my > test system supports four, separate bootable systems (combinations of /boot > partitions and LVM logical volumes for root). > > OK, boot it up and start configuring it, updating (I did a DVD-only install), > etc. Well, one of the things this system does is run qemu-kvm so I had > selected Virtualization during the install. My guest disk images are all on a > separate logical volume which I mount on /.var/lib/libvirt/images but the > system definition/configuration files are separate. When I went to copy them > from the old system I found the was no place to copy them to ... /etc/libvirt/ > was missing. Further checking showed that the libvirt rpm was not installed > ... this is easy enought to fix but it should not have happened ... it should > have been installed when I selected virtualization. Oops, indeed. I've filed this here: https://bugzilla.redhat.com/530015 Thanks for finding this, it's a very serious issue that could have crept by unnoticed. Cheers, Mark. From gianluca.cecchi at gmail.com Wed Oct 21 09:32:40 2009 From: gianluca.cecchi at gmail.com (Gianluca Cecchi) Date: Wed, 21 Oct 2009 11:32:40 +0200 Subject: [fedora-virt] f12 guest memory problems? Message-ID: <561c252c0910210232y6d3fff4bkcfcd0641fc84e809@mail.gmail.com> Hello, F11 x86_64 + rawvirt with latest updates I have a f12 near beta x86_64 kvm guest where now I'm doing "yum update" in it and basically I'm only in gui session without doing anything else Some strange things: - in virt-manager set 400Mb for VM (both the allocated, change, max allocation values... btw how to use them???) my guest sees only about 330Mb [gcecchi at localhost ~]$ free total used free shared buffers cached Mem: 344212 252944 91268 0 4268 26908 -/+ buffers/cache: 221768 122444 Swap: 950264 179092 771172 With dmesg on guest I see something like this: Memory: 329452k/409536k ...... 79696k reserved The 79696k reserved seems to fill the gap... is this a feature in f12 or based on it being a guest? - As you see, the guest is swapping about 180Mb, and actually it is very slowly responsive. Shouldn't it use some of buffers/cache before swapping so much? Is this expected during a yum update (only 34 packages for an overall of 72MB)? Or problem with F12 itself or virtualization layer? Thanks Gianluca -------------- next part -------------- An HTML attachment was scrubbed... URL: From rjones at redhat.com Wed Oct 21 10:28:24 2009 From: rjones at redhat.com (Richard W.M. Jones) Date: Wed, 21 Oct 2009 11:28:24 +0100 Subject: [fedora-virt] virtio block driver performance versus ide In-Reply-To: <1255636225.29670.40.camel@tofu.lib.ucdavis.edu> References: <1255636225.29670.40.camel@tofu.lib.ucdavis.edu> Message-ID: <20091021102824.GA4897@amd.home.annexia.org> Could it be this? https://bugzilla.redhat.com/show_bug.cgi?id=509383 There's a simple adjustment you can try in comment 3. Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones virt-p2v converts physical machines to virtual machines. Boot with a live CD or over the network (PXE) and turn machines into Xen guests. http://et.redhat.com/~rjones/virt-p2v From kenni at kelu.dk Wed Oct 21 11:47:52 2009 From: kenni at kelu.dk (Kenni Lund) Date: Wed, 21 Oct 2009 13:47:52 +0200 Subject: [fedora-virt] virbr0 messing with iptables rules? Message-ID: <3b1f68ef0910210447q66a33617t727e938fe3dff04@mail.gmail.com> Hi I just did a full system update on my F11 server, but after a reboot, new rules were appended to my iptables setup. The iptables and ip6tables services are both disabled: # chkconfig |grep ip.*tables ip6tables 0:off 1:off 2:off 3:off 4:off 5:off 6:off iptables 0:off 1:off 2:off 3:off 4:off 5:off 6:off I set my iptables rules in a custom firewall script in /etc/rc.local, which starts out by flushing all rules. Eg. if I run the script manually after boot, it will "fix" the issue. The extra rules appended are: ----- Chain INPUT (policy DROP) target prot opt source destination ACCEPT udp -- anywhere anywhere udp dpt:domain ACCEPT tcp -- anywhere anywhere tcp dpt:domain ACCEPT udp -- anywhere anywhere udp dpt:bootps ACCEPT tcp -- anywhere anywhere tcp dpt:bootps Chain FORWARD (policy DROP) target prot opt source destination ACCEPT all -- anywhere 192.168.122.0/24 state RELATED,ESTABLISHED ACCEPT all -- 192.168.122.0/24 anywhere ACCEPT all -- anywhere anywhere REJECT all -- anywhere anywhere reject-with icmp-port-unreachable REJECT all -- anywhere anywhere reject-with icmp-port-unreachable ----- I suppose that these rules are related to libvirt, since I have a virbr0 interface with the IP-address 192.168.122.1. Apparently the update changed something, so these firewall rules are appended after /etc/rc.local runs my custom firewall script. What is the correct solution to this? In general, isn't it a bad design decision to have a service mess with the iptables rules, instead of doing this through the iptables/ip6tables services? I would not expect other services to mess with my rules, when I explicitly disabled the build-in iptables services. I only use bridged networking for my virtual machines, so I don't suppose that I need the virbr0 interface after all. Can I disable it somewhere? Best Regards Kenni Lund From rjones at redhat.com Wed Oct 21 15:01:35 2009 From: rjones at redhat.com (Richard W.M. Jones) Date: Wed, 21 Oct 2009 16:01:35 +0100 Subject: [fedora-virt] virtio block driver performance versus ide In-Reply-To: <6172c17e0910210758j131b67ffrb7d09b08d9f06fec@mail.gmail.com> References: <1255636225.29670.40.camel@tofu.lib.ucdavis.edu> <20091021102824.GA4897@amd.home.annexia.org> <6172c17e0910210758j131b67ffrb7d09b08d9f06fec@mail.gmail.com> Message-ID: <20091021150135.GC3408@amd.home.annexia.org> On Wed, Oct 21, 2009 at 08:58:10AM -0600, Thomas S Hatch wrote: > Wow, the rotational change greatly speeds up my vms, as much as 15 fold. > > At the risk of sounding like a complete n00b, is there a way to persistently > change the settings in /sys? I am of course familiar with the sysctl.conf > for /proc/sys. Or is this something that should be changed at the driver > level? It has apparently already been changed (see the new kernels in that Bugzilla entry). In any case, you can't make the change permanent, but you can add the commands to /etc/rc.local so they get executed at each guest boot which has just about the same effect. Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones libguestfs lets you edit virtual machines. Supports shell scripting, bindings from many languages. http://et.redhat.com/~rjones/libguestfs/ See what it can do: http://et.redhat.com/~rjones/libguestfs/recipes.html From psj at familyjenner.co.uk Wed Oct 21 19:28:56 2009 From: psj at familyjenner.co.uk (Paul Jenner) Date: Wed, 21 Oct 2009 20:28:56 +0100 Subject: [fedora-virt] virt-manager and ssh with non root user In-Reply-To: <561c252c0910200826l6d338f22tdd58360d7d3fec08@mail.gmail.com> References: <561c252c0910200826l6d338f22tdd58360d7d3fec08@mail.gmail.com> Message-ID: <1256153337.2230.8.camel@82-71-88-77.dsl.in-addr.zen.co.uk> Hi Gianluca. Only a response to part of your problem but... On Tue, 2009-10-20 at 17:26 +0200, Gianluca Cecchi wrote: > Also, as in virt-manager this connection fails to create (root > prevented from ssh directly as said above), I would like to delete > it... but I didn't find an option for this; > In virt-manager only "connect" and "details" options are available.... Do you mean that you right-click the connection in virt-manager and you only see "Connect" and "Details" in the context menu - no obvious way to delete the connection? If so, try left-clicking to select the connection and then choosing Edit | Delete from the main menu. Does that help? I am not sure if that is what you meant but I keep meaning to raise a UI bug about the differences between the operations available from the right-click context menus and the main menus. Paul -- Paul Jenner From gianluca.cecchi at gmail.com Wed Oct 21 19:43:50 2009 From: gianluca.cecchi at gmail.com (Gianluca Cecchi) Date: Wed, 21 Oct 2009 21:43:50 +0200 Subject: [fedora-virt] virt-manager and ssh with non root user In-Reply-To: <1256153337.2230.8.camel@82-71-88-77.dsl.in-addr.zen.co.uk> References: <561c252c0910200826l6d338f22tdd58360d7d3fec08@mail.gmail.com> <1256153337.2230.8.camel@82-71-88-77.dsl.in-addr.zen.co.uk> Message-ID: <561c252c0910211243l5957c4fdy60589e56fd9b6b79@mail.gmail.com> On Wed, Oct 21, 2009 at 9:28 PM, Paul Jenner wrote: > Hi Gianluca. > > Only a response to part of your problem but... > > On Tue, 2009-10-20 at 17:26 +0200, Gianluca Cecchi wrote: > > Also, as in virt-manager this connection fails to create (root > > prevented from ssh directly as said above), I would like to delete > > it... but I didn't find an option for this; > > In virt-manager only "connect" and "details" options are available.... > > Do you mean that you right-click the connection in virt-manager and you > only see "Connect" and "Details" in the context menu - no obvious way to > delete the connection? > > If so, try left-clicking to select the connection and then choosing Edit > | Delete from the main menu. Does that help? > > I am not sure if that is what you meant but I keep meaning to raise a UI > bug about the differences between the operations available from the > right-click context menus and the main menus. > > Paul > > Hi Paul, yes it was exactly what I meant. I have not at this time ht server at hand. But I tried creating a fake connection form a same way configured server and it works as you said Opened also a bug: https://bugzilla.redhat.com/show_bug.cgi?id=530178 Thanks for your suggestion Gianluca -------------- next part -------------- An HTML attachment was scrubbed... URL: From abo at root.snowtree.se Wed Oct 21 21:10:54 2009 From: abo at root.snowtree.se (Alexander =?ISO-8859-1?Q?Bostr=F6m?=) Date: Wed, 21 Oct 2009 23:10:54 +0200 Subject: [fedora-virt] virt-manager and ssh with non root user In-Reply-To: <561c252c0910200826l6d338f22tdd58360d7d3fec08@mail.gmail.com> References: <561c252c0910200826l6d338f22tdd58360d7d3fec08@mail.gmail.com> Message-ID: <1256159454.24658.11.camel@tempo.alexander.bostrom.net> tis 2009-10-20 klockan 17:26 +0200 skrev Gianluca Cecchi: > Also I cannot export DISPLAY env so I cannot directly ssh to the host > and run its virt-manager program.... This doesn't work? ssh -Xf myrhel54host virt-manager /abo From tirloni at gmail.com Thu Oct 22 01:39:01 2009 From: tirloni at gmail.com (Giovanni Tirloni) Date: Wed, 21 Oct 2009 23:39:01 -0200 Subject: [fedora-virt] State of KVM paravirtualization Message-ID: Hello, I'm trying to understand where KVM stands regarding paravirtualization. I've seen the patches from 2007 by Ingo Molnar (http://people.redhat.com/mingo/kvm-paravirt-patches/) but couldn't find anything newer. Are they integrated to the mainstream kernel? The reason I'm looking at that is because I've a server at home which doesn't have hardware-assisted virtualization support and running qemu in full virtualization mode is a bit slow. I've tried Xen and it sure is fast but I would like to experiment with KVM. Thank you, -- Giovanni. From dlbewley at lib.ucdavis.edu Thu Oct 22 05:13:59 2009 From: dlbewley at lib.ucdavis.edu (Dale Bewley) Date: Wed, 21 Oct 2009 22:13:59 -0700 (PDT) Subject: [fedora-virt] f12-beta geust install fails on f11 In-Reply-To: <621623048.136851256187668894.JavaMail.root@zebra.lib.ucdavis.edu> Message-ID: <928519134.136871256188439896.JavaMail.root@zebra.lib.ucdavis.edu> Host is f11 with virt-manager-0.7.0-7.fc11.x86_64 Using the guest creation wizard, I tried to install f12-beta. I pointed at a http mirror and allowed virt-manager to detect the OS and variant, but it failed with: Unable to complete install: 'OS variant 'fedora12-beta; does not exist in our dictionary for OS type 'linux'' I went back and selected F11, and it works fine so far of course. Would it make sense if the auto-detect could sense alpha and beta releases then punt to a default such as the most recent release of the distribution or some other common denominator? -- Dale Bewley - Unix Administrator - Shields Library - UC Davis GPG: 0xB098A0F3 0D5A 9AEB 43F4 F84C 7EFD 1753 064D 2583 B098 A0F3 From dlaor at redhat.com Thu Oct 22 08:56:01 2009 From: dlaor at redhat.com (Dor Laor) Date: Thu, 22 Oct 2009 10:56:01 +0200 Subject: [fedora-virt] State of KVM paravirtualization In-Reply-To: References: Message-ID: <4AE01E21.9070805@redhat.com> On 10/22/2009 03:39 AM, Giovanni Tirloni wrote: > Hello, > > I'm trying to understand where KVM stands regarding > paravirtualization. I've seen the patches from 2007 by Ingo Molnar > (http://people.redhat.com/mingo/kvm-paravirt-patches/) but couldn't > find anything newer. Are they integrated to the mainstream kernel? The above is not merged and it still relays on hw-assist. KVM has pv drivers, pvclock, and some pv mmu (which under perform the standard vt hw). > > The reason I'm looking at that is because I've a server at home which > doesn't have hardware-assisted virtualization support and running qemu > in full virtualization mode is a bit slow. I've tried Xen and it sure > is fast but I would like to experiment with KVM. There is no plan to support such hardware. Even VMW are moving to this direction. Any new hardware is sold with VT/SVM for the last 2+ years. > > Thank you, > From rvandolson at esri.com Thu Oct 22 12:07:03 2009 From: rvandolson at esri.com (Ray Van Dolson) Date: Thu, 22 Oct 2009 05:07:03 -0700 Subject: [fedora-virt] State of KVM paravirtualization In-Reply-To: References: Message-ID: <20091022120703.GA12655@esri.com> On Wed, Oct 21, 2009 at 06:39:01PM -0700, Giovanni Tirloni wrote: > Hello, > > I'm trying to understand where KVM stands regarding > paravirtualization. I've seen the patches from 2007 by Ingo Molnar > (http://people.redhat.com/mingo/kvm-paravirt-patches/) but couldn't > find anything newer. Are they integrated to the mainstream kernel? > > The reason I'm looking at that is because I've a server at home which > doesn't have hardware-assisted virtualization support and running qemu > in full virtualization mode is a bit slow. I've tried Xen and it sure > is fast but I would like to experiment with KVM. > Fortunately, RH will support Xen at least through RHEL 5's EOL. I'm sure the company I work at is not alone in having a lot of older but still good hardware around that is great for Xen even if it's not cared about by KVM developers so much :) (I'm thinking Dell PE 2850's, etc.. perfectly good servers that these is no reason to toss) OpenVZ might be a good option for you as well depending on your use case. Ray From crobinso at redhat.com Thu Oct 22 12:47:52 2009 From: crobinso at redhat.com (Cole Robinson) Date: Thu, 22 Oct 2009 08:47:52 -0400 Subject: [fedora-virt] f12-beta geust install fails on f11 In-Reply-To: <928519134.136871256188439896.JavaMail.root@zebra.lib.ucdavis.edu> References: <928519134.136871256188439896.JavaMail.root@zebra.lib.ucdavis.edu> Message-ID: <4AE05478.3090901@redhat.com> On 10/22/2009 01:13 AM, Dale Bewley wrote: > Host is f11 with virt-manager-0.7.0-7.fc11.x86_64 > > Using the guest creation wizard, I tried to install f12-beta. I pointed at a http mirror and allowed virt-manager to detect the OS and variant, but it failed with: > > Unable to complete install: 'OS variant 'fedora12-beta; does not exist in our dictionary for OS type 'linux'' > > I went back and selected F11, and it works fine so far of course. Would it make sense if the auto-detect could sense alpha and beta releases then punt to a default such as the most recent release of the distribution or some other common denominator? > Can you try pulling virtinst from updates testing?: yum --enablerepo=updates-testing update python-virtinst There was a relevant fix that I believe hasn't been pushed to stable yet. Please report the results here. Thanks, Cole From dlbewley at lib.ucdavis.edu Thu Oct 22 16:18:30 2009 From: dlbewley at lib.ucdavis.edu (Dale Bewley) Date: Thu, 22 Oct 2009 09:18:30 -0700 (PDT) Subject: [fedora-virt] f12-beta geust install fails on f11 In-Reply-To: <696359488.144791256228285487.JavaMail.root@zebra.lib.ucdavis.edu> Message-ID: <1342460183.144841256228310614.JavaMail.root@zebra.lib.ucdavis.edu> ----- "Cole Robinson" wrote: > On 10/22/2009 01:13 AM, Dale Bewley wrote: > > Host is f11 with virt-manager-0.7.0-7.fc11.x86_64 > > > > Using the guest creation wizard, I tried to install f12-beta. I > pointed at a http mirror and allowed virt-manager to detect the OS and > variant, but it failed with: > > > > Unable to complete install: 'OS variant 'fedora12-beta; does not > exist in our dictionary for OS type 'linux'' > Can you try pulling virtinst from updates testing?: > > yum --enablerepo=updates-testing update python-virtinst > > There was a relevant fix that I believe hasn't been pushed to stable > yet. Please report the results here. > Yes that seems to work. It detects f12-beta as OS Generic. I'll go give it some Bodhi karma. Thanks, Cole. From crobinso at redhat.com Thu Oct 22 16:20:30 2009 From: crobinso at redhat.com (Cole Robinson) Date: Thu, 22 Oct 2009 12:20:30 -0400 Subject: [fedora-virt] f12-beta geust install fails on f11 In-Reply-To: <1342460183.144841256228310614.JavaMail.root@zebra.lib.ucdavis.edu> References: <1342460183.144841256228310614.JavaMail.root@zebra.lib.ucdavis.edu> Message-ID: <4AE0864E.9050308@redhat.com> On 10/22/2009 12:18 PM, Dale Bewley wrote: > ----- "Cole Robinson" wrote: >> On 10/22/2009 01:13 AM, Dale Bewley wrote: >>> Host is f11 with virt-manager-0.7.0-7.fc11.x86_64 >>> >>> Using the guest creation wizard, I tried to install f12-beta. I >> pointed at a http mirror and allowed virt-manager to detect the OS and >> variant, but it failed with: >>> >>> Unable to complete install: 'OS variant 'fedora12-beta; does not >> exist in our dictionary for OS type 'linux'' > >> Can you try pulling virtinst from updates testing?: >> >> yum --enablerepo=updates-testing update python-virtinst >> >> There was a relevant fix that I believe hasn't been pushed to stable >> yet. Please report the results here. >> > > Yes that seems to work. It detects f12-beta as OS Generic. > I'll go give it some Bodhi karma. > > Thanks, Cole. I actually just pushed the latest package to stable, but thanks for the intended karma :) - Cole From dlaor at redhat.com Thu Oct 22 23:00:44 2009 From: dlaor at redhat.com (Dor Laor) Date: Fri, 23 Oct 2009 01:00:44 +0200 Subject: [fedora-virt] State of KVM paravirtualization In-Reply-To: <20091022120703.GA12655@esri.com> References: <20091022120703.GA12655@esri.com> Message-ID: <4AE0E41C.2080303@redhat.com> On 10/22/2009 02:07 PM, Ray Van Dolson wrote: > On Wed, Oct 21, 2009 at 06:39:01PM -0700, Giovanni Tirloni wrote: >> Hello, >> >> I'm trying to understand where KVM stands regarding >> paravirtualization. I've seen the patches from 2007 by Ingo Molnar >> (http://people.redhat.com/mingo/kvm-paravirt-patches/) but couldn't >> find anything newer. Are they integrated to the mainstream kernel? >> >> The reason I'm looking at that is because I've a server at home which >> doesn't have hardware-assisted virtualization support and running qemu >> in full virtualization mode is a bit slow. I've tried Xen and it sure >> is fast but I would like to experiment with KVM. >> > > Fortunately, RH will support Xen at least through RHEL 5's EOL. I'm > sure the company I work at is not alone in having a lot of older but > still good hardware around that is great for Xen even if it's not cared > about by KVM developers so much :) btw: it's not about care. It's not practical of kvm to implement it. By the time it will work, even your Dell will be old... Also hw-assist virt with nested paging is a killer, it outperforms any PV by any vendor. > > (I'm thinking Dell PE 2850's, etc.. perfectly good servers that these > is no reason to toss) > > OpenVZ might be a good option for you as well depending on your use > case. > > Ray > > _______________________________________________ > Fedora-virt mailing list > Fedora-virt at redhat.com > https://www.redhat.com/mailman/listinfo/fedora-virt From rvandolson at esri.com Thu Oct 22 23:03:40 2009 From: rvandolson at esri.com (Ray Van Dolson) Date: Thu, 22 Oct 2009 16:03:40 -0700 Subject: [fedora-virt] State of KVM paravirtualization In-Reply-To: <4AE0E41C.2080303@redhat.com> References: <20091022120703.GA12655@esri.com> <4AE0E41C.2080303@redhat.com> Message-ID: <20091022230339.GB29420@esri.com> On Thu, Oct 22, 2009 at 04:00:44PM -0700, Dor Laor wrote: > On 10/22/2009 02:07 PM, Ray Van Dolson wrote: > > On Wed, Oct 21, 2009 at 06:39:01PM -0700, Giovanni Tirloni wrote: > >> Hello, > >> > >> I'm trying to understand where KVM stands regarding > >> paravirtualization. I've seen the patches from 2007 by Ingo Molnar > >> (http://people.redhat.com/mingo/kvm-paravirt-patches/) but couldn't > >> find anything newer. Are they integrated to the mainstream kernel? > >> > >> The reason I'm looking at that is because I've a server at home which > >> doesn't have hardware-assisted virtualization support and running qemu > >> in full virtualization mode is a bit slow. I've tried Xen and it sure > >> is fast but I would like to experiment with KVM. > >> > > > > Fortunately, RH will support Xen at least through RHEL 5's EOL. I'm > > sure the company I work at is not alone in having a lot of older but > > still good hardware around that is great for Xen even if it's not cared > > about by KVM developers so much :) > > btw: it's not about care. It's not practical of kvm to implement it. By > the time it will work, even your Dell will be old... Also hw-assist virt > with nested paging is a killer, it outperforms any PV by any vendor. Understood. It's a priority thing either way. Makes more sense to pursue the new hardware than the old. And Xen works OK for the old stuff in the meantime. :) > > > > > (I'm thinking Dell PE 2850's, etc.. perfectly good servers that these > > is no reason to toss) > > > > OpenVZ might be a good option for you as well depending on your use > > case. > > > > Ray From jameshubbard at gmail.com Thu Oct 22 23:19:02 2009 From: jameshubbard at gmail.com (James Hubbard) Date: Thu, 22 Oct 2009 19:19:02 -0400 Subject: [fedora-virt] State of KVM paravirtualization In-Reply-To: <4AE01E21.9070805@redhat.com> References: <4AE01E21.9070805@redhat.com> Message-ID: On Thu, Oct 22, 2009 at 4:56 AM, Dor Laor wrote: > There is no plan to support such hardware. Even VMW are moving to this > direction. Any new hardware is sold with VT/SVM for the last 2+ years. That comment is true for AMD. However, even some of the higher end Intel chips might not support VT. http://arstechnica.com/microsoft/news/2009/05/r2e-microsoft-intel-goof-up-windows-7s-xp-mode.ars From markmc at redhat.com Fri Oct 23 11:04:09 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Fri, 23 Oct 2009 12:04:09 +0100 Subject: [fedora-virt] virbr0 messing with iptables rules? In-Reply-To: <3b1f68ef0910210447q66a33617t727e938fe3dff04@mail.gmail.com> References: <3b1f68ef0910210447q66a33617t727e938fe3dff04@mail.gmail.com> Message-ID: <1256295849.31881.66.camel@blaa> On Wed, 2009-10-21 at 13:47 +0200, Kenni Lund wrote: > Hi > > I just did a full system update on my F11 server, but after a reboot, > new rules were appended to my iptables setup. > > The iptables and ip6tables services are both disabled: > # chkconfig |grep ip.*tables > ip6tables 0:off 1:off 2:off 3:off 4:off 5:off 6:off > iptables 0:off 1:off 2:off 3:off 4:off 5:off 6:off > > I set my iptables rules in a custom firewall script in /etc/rc.local, > which starts out by flushing all rules. > Eg. if I run the script manually after boot, it will "fix" the issue. > > The extra rules appended are: > ----- > Chain INPUT (policy DROP) > target prot opt source destination > ACCEPT udp -- anywhere anywhere udp > dpt:domain > ACCEPT tcp -- anywhere anywhere tcp > dpt:domain > ACCEPT udp -- anywhere anywhere udp > dpt:bootps > ACCEPT tcp -- anywhere anywhere tcp dpt:bootps > > Chain FORWARD (policy DROP) > target prot opt source destination > ACCEPT all -- anywhere 192.168.122.0/24 state > RELATED,ESTABLISHED > ACCEPT all -- 192.168.122.0/24 anywhere > ACCEPT all -- anywhere anywhere > REJECT all -- anywhere anywhere > reject-with icmp-port-unreachable > REJECT all -- anywhere anywhere > reject-with icmp-port-unreachable > ----- > > I suppose that these rules are related to libvirt, since I have a > virbr0 interface with the IP-address 192.168.122.1. Apparently the > update changed something, so these firewall rules are appended after > /etc/rc.local runs my custom firewall script. > > What is the correct solution to this? In general, isn't it a bad > design decision to have a service mess with the iptables rules, > instead of doing this through the iptables/ip6tables services? I would > not expect other services to mess with my rules, when I explicitly > disabled the build-in iptables services. Yeah, we're not happy with the way we're integrating with iptables. We proposed one way of doing it and tried out the counter-proposal, but we're back to square one again. The whole saga is documented here: https://bugzilla.redhat.com/227011 > I only use bridged networking for my virtual machines, so I don't > suppose that I need the virbr0 interface after all. Can I disable it > somewhere? This should do it: $> virsh net-destroy default $> virsh net-autostart --disable default Cheers, Mark. From hughsient at gmail.com Fri Oct 23 13:14:49 2009 From: hughsient at gmail.com (Richard Hughes) Date: Fri, 23 Oct 2009 14:14:49 +0100 Subject: [fedora-virt] Using the virtio drivers in Windows XP setup Message-ID: <15e53e180910230614n370f2d72m63232e9acf6d8396@mail.gmail.com> This may be useful to somebody. To use the virtio drivers in windows XP setup, you have to do the following: * create a 1.44Mb image file * mount it by loopback * format it with vfat * copy the Install/Xp/x86/viostor.sys, Install/Xp/x86/wnet.inf, and the attached txtsetup.oem file to the root of the mounted image * umount the loop device * attach the floppy image as a floppy storage element in the VM's details pane * boot the VM * remember to press F6 when booting the windows xp setup and select the VirtoIO device. The txtsetup.oem was just cobbled together after reading various sources online, and hacking values out of the supplied inf file. Richard. p.s. I'm still unable to install XP using ide, scsi or virtio drivers as it gives the message "Setup was unable to format the partition. The disk maybe damaged." -- any ideas welcome. From wildfire at progsoc.org Fri Oct 23 15:22:18 2009 From: wildfire at progsoc.org (Anand Kumria) Date: Fri, 23 Oct 2009 16:22:18 +0100 Subject: [fedora-virt] Using the virtio drivers in Windows XP setup In-Reply-To: <15e53e180910230614n370f2d72m63232e9acf6d8396@mail.gmail.com> References: <15e53e180910230614n370f2d72m63232e9acf6d8396@mail.gmail.com> Message-ID: <971f65790910230822l66c3681cld6b3c8840abdc44d@mail.gmail.com> Hi Richard, The txtsetup.oem file you mentioned was not with your email, I am not sure if you forgot or the mailing list eats attachments. Perhaps you have a URL to the file? It also sounds like a small script would make it easier for everyone to install the drivers. Regards, Anand On Fri, Oct 23, 2009 at 2:14 PM, Richard Hughes wrote: > This may be useful to somebody. To use the virtio drivers in windows > XP setup, you have to do the following: > > * create a 1.44Mb image file > * mount it by loopback > * format it with vfat > * copy the Install/Xp/x86/viostor.sys, Install/Xp/x86/wnet.inf, and > the attached txtsetup.oem file to the root of the mounted image > * umount the loop device > * attach the floppy image as a floppy storage element in the VM's details > pane > * boot the VM > * remember to press F6 when booting the windows xp setup and select > the VirtoIO device. > > The txtsetup.oem was just cobbled together after reading various > sources online, and hacking values out of the supplied inf file. > > Richard. > > p.s. I'm still unable to install XP using ide, scsi or virtio drivers > as it gives the message "Setup was unable to format the partition. The > disk maybe damaged." -- any ideas welcome. > > _______________________________________________ > Fedora-virt mailing list > Fedora-virt at redhat.com > https://www.redhat.com/mailman/listinfo/fedora-virt > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From gene at czarc.net Fri Oct 23 15:26:33 2009 From: gene at czarc.net (Gene Czarcinski) Date: Fri, 23 Oct 2009 11:26:33 -0400 Subject: [fedora-virt] Request for confirmation/assistance Message-ID: <200910231126.33265.gene@czarc.net> I have a qemu-kvm problem which causes a qemu-kvm guest to "suddenly" crash: https://bugzilla.redhat.com/show_bug.cgi?id=528754 This has happened to me at least 5 times on a F12-rawhide host running qemu-kvm-0.11.0-6.fc12.x86_64 The guests (which crashed) have been F11 and F12 with (at the time) all updates applied. The problem appears to be related to "sound" which does not work currently for a guest and which I do not use. I do not know what action on my part "causes" the problem. When the problem occurs, there have been (IIRC) two guests running (F11 and F12). When the problem happens, the guest console suddenly becomes a white screen with "Guest not running". There are messages in /var/log/messages which say that qemu-kvm has had a segfault. OK, has anyone else seen anything like this? So far, there has been insufficient information to give the developers, etc. any idea of where the problem is. I suspect it would be helpful if someone else had the problem. Gene From rjones at redhat.com Fri Oct 23 15:44:37 2009 From: rjones at redhat.com (Richard W.M. Jones) Date: Fri, 23 Oct 2009 16:44:37 +0100 Subject: [fedora-virt] Using the virtio drivers in Windows XP setup In-Reply-To: <971f65790910230822l66c3681cld6b3c8840abdc44d@mail.gmail.com> References: <15e53e180910230614n370f2d72m63232e9acf6d8396@mail.gmail.com> <971f65790910230822l66c3681cld6b3c8840abdc44d@mail.gmail.com> Message-ID: <20091023154437.GB25184@amd.home.annexia.org> On Fri, Oct 23, 2009 at 04:22:18PM +0100, Anand Kumria wrote: > Hi Richard, > > The txtsetup.oem file you mentioned was not with your email, I am not sure > if you forgot or the mailing list eats attachments. > > Perhaps you have a URL to the file? It also sounds like a small script would > make it easier for everyone to install the drivers. I asked Richard and he sent me it: ---------cut----------- [Disks] d1 = "Viostor SCSI driver disk",\disk1.tag,\ [Defaults] SCSI = viostor [SCSI] viostor = "Viostor SCSI Controller" [Files.SCSI.viostor] driver = d1,viostor.sys,viostor inf = d1,wnet.inf [HardwareIds.scsi.viostor] id = "PCI\VEN_1AF4&DEV_1001","viostor" ---------cut----------- Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones Read my programming blog: http://rwmj.wordpress.com Fedora now supports 80 OCaml packages (the OPEN alternative to F#) http://cocan.org/getting_started_with_ocaml_on_red_hat_and_fedora From jonr at destar.net Fri Oct 23 17:11:12 2009 From: jonr at destar.net (jonr at destar.net) Date: Fri, 23 Oct 2009 09:11:12 -0800 Subject: [fedora-virt] Using the virtio drivers in Windows XP setup In-Reply-To: <15e53e180910230614n370f2d72m63232e9acf6d8396@mail.gmail.com> References: <15e53e180910230614n370f2d72m63232e9acf6d8396@mail.gmail.com> Message-ID: <20091023091112.19394pt2tws6x9c0@www.destar.net> Quoting Richard Hughes : > p.s. I'm still unable to install XP using ide, scsi or virtio drivers > as it gives the message "Setup was unable to format the partition. The > disk maybe damaged." -- any ideas welcome. > Can you set the disk up beforehand and then tell it not to partition it but use whats already there? Jon From pasik at iki.fi Mon Oct 26 08:58:12 2009 From: pasik at iki.fi (Pasi =?iso-8859-1?Q?K=E4rkk=E4inen?=) Date: Mon, 26 Oct 2009 10:58:12 +0200 Subject: [fedora-virt] State of KVM paravirtualization In-Reply-To: <4AE0E41C.2080303@redhat.com> References: <20091022120703.GA12655@esri.com> <4AE0E41C.2080303@redhat.com> Message-ID: <20091026085812.GY1434@reaktio.net> On Fri, Oct 23, 2009 at 01:00:44AM +0200, Dor Laor wrote: > On 10/22/2009 02:07 PM, Ray Van Dolson wrote: > >On Wed, Oct 21, 2009 at 06:39:01PM -0700, Giovanni Tirloni wrote: > >>Hello, > >> > >> I'm trying to understand where KVM stands regarding > >>paravirtualization. I've seen the patches from 2007 by Ingo Molnar > >>(http://people.redhat.com/mingo/kvm-paravirt-patches/) but couldn't > >>find anything newer. Are they integrated to the mainstream kernel? > >> > >> The reason I'm looking at that is because I've a server at home which > >>doesn't have hardware-assisted virtualization support and running qemu > >>in full virtualization mode is a bit slow. I've tried Xen and it sure > >>is fast but I would like to experiment with KVM. > >> > > > >Fortunately, RH will support Xen at least through RHEL 5's EOL. I'm > >sure the company I work at is not alone in having a lot of older but > >still good hardware around that is great for Xen even if it's not cared > >about by KVM developers so much :) > > btw: it's not about care. It's not practical of kvm to implement it. By > the time it will work, even your Dell will be old... Also hw-assist virt > with nested paging is a killer, it outperforms any PV by any vendor. > As I've understood this is the case with KVM, since pv-mmu stuff in KVM isn't as fast as Xen has. ie. with Xen PV is still faster than HVM (hw-assisted), at least for some worksloads. HVM (hw-assisted paging) has other/different performance hits. -- Pasi From gianluca.cecchi at gmail.com Mon Oct 26 10:10:16 2009 From: gianluca.cecchi at gmail.com (Gianluca Cecchi) Date: Mon, 26 Oct 2009 11:10:16 +0100 Subject: [fedora-virt] Request for confirmation/assistance Message-ID: <561c252c0910260310v4aebe67ah2b0854bd990d3336@mail.gmail.com> On Fri, 23 Oct 2009 11:26:33 -0400 Gene Czarcinski wrote: > The problem appears to be related to "sound" which does not work currently for > a guest and which I do not use. Hello, sorry Gene, can I jump in and ask for having a way to disable sound (aka not having a sound card through the hardware pool components) for a guest during its creation/installation? When you use Qemu/KVM to host several servers, I think that usually you don't want sound at all..... Thanks and let me know if I have to post a bug as enhancement.... if this is not already planned as a feature. Gianluca -------------- next part -------------- An HTML attachment was scrubbed... URL: From giallu at gmail.com Mon Oct 26 22:29:08 2009 From: giallu at gmail.com (Gianluca Sforna) Date: Mon, 26 Oct 2009 23:29:08 +0100 Subject: [fedora-virt] Porting existing virtual machines to newer Fedora Message-ID: I'm installing Fedora 12 (full install, no upgrade) on my laptop where I have some virtual machines configured via F11's virt-manager. After installation I'd like to recover from a backup the existing virtual machines; is there any documentation on the proper/recommended/safer method for doing the job? TIA G. -- Gianluca Sforna http://morefedora.blogspot.com http://www.linkedin.com/in/gianlucasforna From tom.horsley at att.net Mon Oct 26 23:11:06 2009 From: tom.horsley at att.net (Tom Horsley) Date: Mon, 26 Oct 2009 19:11:06 -0400 Subject: [fedora-virt] Porting existing virtual machines to newer Fedora In-Reply-To: References: Message-ID: <20091026191106.4f3a44ad@zooty> On Mon, 26 Oct 2009 23:29:08 +0100 Gianluca Sforna wrote: > After installation I'd like to recover from a backup the existing > virtual machines; is there any documentation on the > proper/recommended/safer method for doing the job? What has worked for me is doing "virsh dumpxml" to save the machine def on f11, then "virsh define" to reload the xml on f12. (As long as libvirt is configured the same way first, which for me means getting my bridge defined for bridge networking). From ieidus at redhat.com Mon Oct 26 22:55:06 2009 From: ieidus at redhat.com (Izik Eidus) Date: Tue, 27 Oct 2009 00:55:06 +0200 Subject: [fedora-virt] disabling ksm by default Message-ID: <4AE628CA.1010807@redhat.com> Hi, I saw that Fedora 12 will have ksm tunning script that control the ksm speed / kernel pages allocation. The only problem that I have is - as far as I remember ksm is by default enabled and only with 2000 kernel pages... What I am worried about is that users wont use the ksm tunning script and would just run ksm with this 2000 kernel pages - the result would be that ksm will probably merge just the zero pages (that could be alot of memory) and the user might not know that much more memory can be saved... Is it possible to at least make ksm disabled by default? so the users will have to run the ksm tunning script when they want to start ksm? And if we set it to disabled by default, cant we set the initialized values into more realistic value? (like 1/4 of the memory in the mainline kernels? ) Thanks. From markmc at redhat.com Tue Oct 27 10:08:49 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Tue, 27 Oct 2009 10:08:49 +0000 Subject: [fedora-virt] Porting existing virtual machines to newer Fedora In-Reply-To: References: Message-ID: <1256638129.25282.57.camel@blaa> On Mon, 2009-10-26 at 23:29 +0100, Gianluca Sforna wrote: > I'm installing Fedora 12 (full install, no upgrade) on my laptop where > I have some virtual machines configured via F11's virt-manager. > > After installation I'd like to recover from a backup the existing > virtual machines; is there any documentation on the > proper/recommended/safer method for doing the job? Not sure if we have any nice documentation on this, but it should simply be a matter of backing up and restoring the guest's XML configuration and disk images. Cheers, Mark. From markmc at redhat.com Tue Oct 27 10:12:49 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Tue, 27 Oct 2009 10:12:49 +0000 Subject: [fedora-virt] disabling ksm by default In-Reply-To: <4AE628CA.1010807@redhat.com> References: <4AE628CA.1010807@redhat.com> Message-ID: <1256638369.25282.60.camel@blaa> Hi Izik, On Tue, 2009-10-27 at 00:55 +0200, Izik Eidus wrote: > Hi, I saw that Fedora 12 will have ksm tunning script that control the > ksm speed / kernel pages allocation. > > The only problem that I have is - as far as I remember ksm is by default > enabled and only with 2000 kernel pages... > What I am worried about is that users wont use the ksm tunning script > and would just run ksm with this 2000 kernel pages - > the result would be that ksm will probably merge just the zero pages > (that could be alot of memory) and the user might not know > that much more memory can be saved... > > Is it possible to at least make ksm disabled by default? so the users > will have to run the ksm tunning script when they want to start ksm? ksm is disabled by default currently > And if we set it to disabled by default, cant we set the initialized > values into more realistic value? (like 1/4 of the memory in the > mainline kernels? ) Here's the logic we have in the init script: # unless KSM_MAX_KERNEL_PAGES is set, let ksm munch up to half of total memory. default_max_kernel_pages () { local total pagesize total=`awk '/^MemTotal:/ {print $2}' /proc/meminfo` pagesize=`getconf PAGESIZE` echo $[total * 1024 / pagesize / 2] } git tree for the scripts here: http://gitorious.org/ksm-control-scripts/ksm-control-scripts Suggestions, comments, patches etc. welcome Thanks, Mark. From ieidus at redhat.com Tue Oct 27 10:24:13 2009 From: ieidus at redhat.com (Izik Eidus) Date: Tue, 27 Oct 2009 12:24:13 +0200 Subject: [fedora-virt] disabling ksm by default In-Reply-To: <1256638369.25282.60.camel@blaa> References: <4AE628CA.1010807@redhat.com> <1256638369.25282.60.camel@blaa> Message-ID: <4AE6CA4D.9010604@redhat.com> On 10/27/2009 12:12 PM, Mark McLoughlin wrote: > Hi Izik, > > On Tue, 2009-10-27 at 00:55 +0200, Izik Eidus wrote: > >> Hi, I saw that Fedora 12 will have ksm tunning script that control the >> ksm speed / kernel pages allocation. >> >> The only problem that I have is - as far as I remember ksm is by default >> enabled and only with 2000 kernel pages... >> What I am worried about is that users wont use the ksm tunning script >> and would just run ksm with this 2000 kernel pages - >> the result would be that ksm will probably merge just the zero pages >> (that could be alot of memory) and the user might not know >> that much more memory can be saved... >> >> Is it possible to at least make ksm disabled by default? so the users >> will have to run the ksm tunning script when they want to start ksm? >> > ksm is disabled by default currently > Ok so we are fine then. > >> And if we set it to disabled by default, cant we set the initialized >> values into more realistic value? (like 1/4 of the memory in the >> mainline kernels? ) >> > Here's the logic we have in the init script: > > # unless KSM_MAX_KERNEL_PAGES is set, let ksm munch up to half of total memory. > default_max_kernel_pages () { > local total pagesize > total=`awk '/^MemTotal:/ {print $2}' /proc/meminfo` > pagesize=`getconf PAGESIZE` > echo $[total * 1024 / pagesize / 2] > } > > git tree for the scripts here: > > http://gitorious.org/ksm-control-scripts/ksm-control-scripts > > Suggestions, comments, patches etc. welcome > > Thanks. From danken at redhat.com Tue Oct 27 12:37:50 2009 From: danken at redhat.com (Dan Kenigsberg) Date: Tue, 27 Oct 2009 14:37:50 +0200 Subject: [fedora-virt] disabling ksm by default In-Reply-To: <4AE6CA4D.9010604@redhat.com> References: <4AE628CA.1010807@redhat.com> <1256638369.25282.60.camel@blaa> <4AE6CA4D.9010604@redhat.com> Message-ID: <20091027123749.GA4530@redhat.com> On Tue, Oct 27, 2009 at 12:24:13PM +0200, Izik Eidus wrote: > On 10/27/2009 12:12 PM, Mark McLoughlin wrote: >> Hi Izik, >> >> On Tue, 2009-10-27 at 00:55 +0200, Izik Eidus wrote: >> >>> Hi, I saw that Fedora 12 will have ksm tunning script that control the >>> ksm speed / kernel pages allocation. >>> >>> The only problem that I have is - as far as I remember ksm is by default >>> enabled and only with 2000 kernel pages... >>> What I am worried about is that users wont use the ksm tunning script >>> and would just run ksm with this 2000 kernel pages - >>> the result would be that ksm will probably merge just the zero pages >>> (that could be alot of memory) and the user might not know >>> that much more memory can be saved... >>> >>> Is it possible to at least make ksm disabled by default? so the users >>> will have to run the ksm tunning script when they want to start ksm? >>> >> ksm is disabled by default currently >> > > Ok so we are fine then. > Actually, Izik is refering to the fact that ksm is, by default, running when the kernel boots. He'd like to see something like diff -u -r1.1 linux-2.6-ksm-updates.patch --- linux-2.6-ksm-updates.patch 7 Aug 2009 19:07:39 -0000 1.1 +++ linux-2.6-ksm-updates.patch 27 Oct 2009 12:34:57 -0000 @@ -1661,7 +1661,7 @@ #define KSM_RUN_MERGE 1 #define KSM_RUN_UNMERGE 2 -static unsigned int ksm_run; -+static unsigned int ksm_run = KSM_RUN_MERGE; ++static unsigned int ksm_run = KSM_RUN_STOP; static DECLARE_WAIT_QUEUE_HEAD(ksm_thread_wait); static DEFINE_MUTEX(ksm_thread_mutex); to behave more like upstream 2.6.32 http://osdir.com/ml/linux-kernel/2009-09/msg09309.html http://osdir.com/ml/linux-kernel/2009-09/msg09777.html From jmforbes at linuxtx.org Tue Oct 27 14:42:09 2009 From: jmforbes at linuxtx.org (Justin M. Forbes) Date: Tue, 27 Oct 2009 09:42:09 -0500 Subject: [fedora-virt] disabling ksm by default In-Reply-To: <20091027123749.GA4530@redhat.com> References: <4AE628CA.1010807@redhat.com> <1256638369.25282.60.camel@blaa> <4AE6CA4D.9010604@redhat.com> <20091027123749.GA4530@redhat.com> Message-ID: <20091027144209.GA13805@linuxtx.org> On Tue, Oct 27, 2009 at 02:37:50PM +0200, Dan Kenigsberg wrote: > On Tue, Oct 27, 2009 at 12:24:13PM +0200, Izik Eidus wrote: > > On 10/27/2009 12:12 PM, Mark McLoughlin wrote: > >> Hi Izik, > >> > >> On Tue, 2009-10-27 at 00:55 +0200, Izik Eidus wrote: > >> > >>> Hi, I saw that Fedora 12 will have ksm tunning script that control the > >>> ksm speed / kernel pages allocation. There are actually 2 init scripts. One to turn on ksm, and one for tuning. The actual ksm init script simply makes sure ksm is turned on in the kernel and sets max_kernel_pages to half of system memory. The ksmtuned script is a bit more involved. > >>> The only problem that I have is - as far as I remember ksm is by default > >>> enabled and only with 2000 kernel pages... > >>> What I am worried about is that users wont use the ksm tunning script > >>> and would just run ksm with this 2000 kernel pages - > >>> the result would be that ksm will probably merge just the zero pages > >>> (that could be alot of memory) and the user might not know > >>> that much more memory can be saved... > >>> > >>> Is it possible to at least make ksm disabled by default? so the users > >>> will have to run the ksm tunning script when they want to start ksm? > >>> > > Actually, Izik is refering to the fact that ksm is, by default, running > when the kernel boots. > While I do understand what you are saying, I don't think it is worth making a kernel change for at this point in the cycle. Because ksm itself has a separate initscript, people who wish to use ksm will likely turn it on. This sets the max_kernel_pages to a reasonable value. People who are not interested enough to turn on the ksm service are probably not the kind of people who will be checking to see how effective ksm is at all. Justin From ieidus at redhat.com Tue Oct 27 15:31:48 2009 From: ieidus at redhat.com (Izik Eidus) Date: Tue, 27 Oct 2009 17:31:48 +0200 Subject: [fedora-virt] disabling ksm by default In-Reply-To: <20091027144209.GA13805@linuxtx.org> References: <4AE628CA.1010807@redhat.com> <1256638369.25282.60.camel@blaa> <4AE6CA4D.9010604@redhat.com> <20091027123749.GA4530@redhat.com> <20091027144209.GA13805@linuxtx.org> Message-ID: <4AE71264.7000102@redhat.com> On 10/27/2009 04:42 PM, Justin M. Forbes wrote: > On Tue, Oct 27, 2009 at 02:37:50PM +0200, Dan Kenigsberg wrote: > >> On Tue, Oct 27, 2009 at 12:24:13PM +0200, Izik Eidus wrote: >> >>> On 10/27/2009 12:12 PM, Mark McLoughlin wrote: >>> >>>> Hi Izik, >>>> >>>> On Tue, 2009-10-27 at 00:55 +0200, Izik Eidus wrote: >>>> >>>> >>>>> Hi, I saw that Fedora 12 will have ksm tunning script that control the >>>>> ksm speed / kernel pages allocation. >>>>> > There are actually 2 init scripts. One to turn on ksm, and one for tuning. > The actual ksm init script simply makes sure ksm is turned on in the kernel > and sets max_kernel_pages to half of system memory. The ksmtuned script is > a bit more involved. > > >>>>> The only problem that I have is - as far as I remember ksm is by default >>>>> enabled and only with 2000 kernel pages... >>>>> What I am worried about is that users wont use the ksm tunning script >>>>> and would just run ksm with this 2000 kernel pages - >>>>> the result would be that ksm will probably merge just the zero pages >>>>> (that could be alot of memory) and the user might not know >>>>> that much more memory can be saved... >>>>> >>>>> Is it possible to at least make ksm disabled by default? so the users >>>>> will have to run the ksm tunning script when they want to start ksm? >>>>> >>>>> >> Actually, Izik is refering to the fact that ksm is, by default, running >> when the kernel boots. >> >> > While I do understand what you are saying, I don't think it is worth making > a kernel change for at this point in the cycle. Because ksm itself has a > separate initscript, people who wish to use ksm will likely turn it on. > This sets the max_kernel_pages to a reasonable value. People who are not > interested enough to turn on the ksm service are probably not the kind of > people who will be checking to see how effective ksm is at all. > > Justin > from http://forum.zwame.pt/showthread.php?p=5449039: "I tested KSM out by creating a couple of Ubuntu 9.04 VMs with 1GB of RAM apiece and a Windows 7 VM with 2GB of RAM. Together, these VMs laid claim to the bulk of the 4GB of RAM available on my test Fedora 12 system. When I switched KSM on, I watched the memory usage on my test machine fall, fairly quickly, from 3.1GB to 2.1GB as my system identified and merged duplicate memory pages. I want to see KSM in action on a more realistically outfitted system, but I'm impressed the capability as I've seen it so far. Beyond KSM, I'm pleased to see that in Fedora 12, KVM will support hotplugging for virtual network adapters, and will present guest machines with an emulated hardware platform that remains consistent across upgrades of the hypervisor. Linux OSes tend not to care when hardware is changed underneath them, but this can cause problems with Windows. I've experienced broken Windows VM installs following KVM upgrades, and I welcome this improvement." To me it sound that users have no idea about this script, and ksm merge to him just the zero pages of windows 7... My feeling is that 99% of the ppl in the world that will use it, would just see the zero page merged and think "that is it..." The current behaivor in fedora 12 is misleading the user (at least it seems to me that it misslead that specific user) Btw we can set this value from userspace if we want "echo 0 > /sys/kernel/mm/ksm/run" From jmforbes at linuxtx.org Tue Oct 27 15:46:33 2009 From: jmforbes at linuxtx.org (Justin M. Forbes) Date: Tue, 27 Oct 2009 10:46:33 -0500 Subject: [fedora-virt] disabling ksm by default In-Reply-To: <4AE71264.7000102@redhat.com> References: <4AE628CA.1010807@redhat.com> <1256638369.25282.60.camel@blaa> <4AE6CA4D.9010604@redhat.com> <20091027123749.GA4530@redhat.com> <20091027144209.GA13805@linuxtx.org> <4AE71264.7000102@redhat.com> Message-ID: <1256658393.2803.31.camel@fedora64.linuxtx.org> On Tue, 2009-10-27 at 17:31 +0200, Izik Eidus wrote: > On 10/27/2009 04:42 PM, Justin M. Forbes wrote: > > > While I do understand what you are saying, I don't think it is worth making > > a kernel change for at this point in the cycle. Because ksm itself has a > > separate initscript, people who wish to use ksm will likely turn it on. > > This sets the max_kernel_pages to a reasonable value. People who are not > > interested enough to turn on the ksm service are probably not the kind of > > people who will be checking to see how effective ksm is at all. > > > To me it sound that users have no idea about this script, and ksm merge > to him just the zero pages of windows 7... > > My feeling is that 99% of the ppl in the world that will use it, would > just see the zero page merged and think "that is it..." > The current behaivor in fedora 12 is misleading the user (at least it > seems to me that it misslead that specific user) > > Btw we can set this value from userspace if we want "echo 0 > > /sys/kernel/mm/ksm/run" Where would we put that in userspace? Another init script? I just updated to documentation on the feature page (and hopefully in the final release notes) to say that ksm can be enabled by running 'sudo chkconfig ksm on' in effort to draw more attention to these scripts. The other option is to make the ksm init script default to on, and just leave ksmtuned off at system start. Since ksm is on by the default kernel, all we would be doing here is changing max_kernel_pages to a machine specific value. The ksm init script is part of the qemu package, so it would only be changed on systems doing virt anyway. Justin From ewan at macmahon.me.uk Tue Oct 27 15:50:16 2009 From: ewan at macmahon.me.uk (Ewan Mac Mahon) Date: Tue, 27 Oct 2009 15:50:16 +0000 Subject: [fedora-virt] disabling ksm by default In-Reply-To: <1256658393.2803.31.camel@fedora64.linuxtx.org> References: <4AE628CA.1010807@redhat.com> <1256638369.25282.60.camel@blaa> <4AE6CA4D.9010604@redhat.com> <20091027123749.GA4530@redhat.com> <20091027144209.GA13805@linuxtx.org> <4AE71264.7000102@redhat.com> <1256658393.2803.31.camel@fedora64.linuxtx.org> Message-ID: <20091027155015.GE3591@macmahon.me.uk> On Tue, Oct 27, 2009 at 10:46:33AM -0500, Justin M. Forbes wrote: > The other option is to make the ksm init script default to on, and just > leave ksmtuned off at system start. I'm sure this is a silly question, but what's the downside to just turning them both on by default? Ewan -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 197 bytes Desc: not available URL: From markmc at redhat.com Tue Oct 27 15:55:16 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Tue, 27 Oct 2009 15:55:16 +0000 Subject: [fedora-virt] disabling ksm by default In-Reply-To: <1256658393.2803.31.camel@fedora64.linuxtx.org> References: <4AE628CA.1010807@redhat.com> <1256638369.25282.60.camel@blaa> <4AE6CA4D.9010604@redhat.com> <20091027123749.GA4530@redhat.com> <20091027144209.GA13805@linuxtx.org> <4AE71264.7000102@redhat.com> <1256658393.2803.31.camel@fedora64.linuxtx.org> Message-ID: <1256658916.25282.75.camel@blaa> On Tue, 2009-10-27 at 10:46 -0500, Justin M. Forbes wrote: > On Tue, 2009-10-27 at 17:31 +0200, Izik Eidus wrote: > > On 10/27/2009 04:42 PM, Justin M. Forbes wrote: > > > > > While I do understand what you are saying, I don't think it is worth making > > > a kernel change for at this point in the cycle. Because ksm itself has a > > > separate initscript, people who wish to use ksm will likely turn it on. > > > This sets the max_kernel_pages to a reasonable value. People who are not > > > interested enough to turn on the ksm service are probably not the kind of > > > people who will be checking to see how effective ksm is at all. > > > > > To me it sound that users have no idea about this script, and ksm merge > > to him just the zero pages of windows 7... > > > > My feeling is that 99% of the ppl in the world that will use it, would > > just see the zero page merged and think "that is it..." > > The current behaivor in fedora 12 is misleading the user (at least it > > seems to me that it misslead that specific user) > > > > Btw we can set this value from userspace if we want "echo 0 > > > /sys/kernel/mm/ksm/run" > > Where would we put that in userspace? Another init script? I just > updated to documentation on the feature page (and hopefully in the final > release notes) to say that ksm can be enabled by running 'sudo chkconfig > ksm on' in effort to draw more attention to these scripts. > > The other option is to make the ksm init script default to on, and just > leave ksmtuned off at system start. Since ksm is on by the default > kernel, all we would be doing here is changing max_kernel_pages to a > machine specific value. The ksm init script is part of the qemu > package, so it would only be changed on systems doing virt anyway. For Fedora 13, it'll be off by default in the kernel and the recommended way of switching it on is with 'chkconfig ksm on' For Fedora 12, it's on by default in the kernel, 'chkconfig ksm on' just changes max pages and the only way of disabling it is by manually writing zero to /sys/kernel/mm/ksm/run IMHO, if we are to change anything for Fedora 12, we should move closer to Fedora 13 behaviour, not further away from it - i.e. make it off by default in the kernel, but I'm fine with delaying that until post GA We'll probably end up with this behaviour in F12 updates at some point anyway when 2.6.32 is pulled in Cheers, Mark. From ieidus at redhat.com Tue Oct 27 15:57:52 2009 From: ieidus at redhat.com (Izik Eidus) Date: Tue, 27 Oct 2009 17:57:52 +0200 Subject: [fedora-virt] disabling ksm by default In-Reply-To: <1256658393.2803.31.camel@fedora64.linuxtx.org> References: <4AE628CA.1010807@redhat.com> <1256638369.25282.60.camel@blaa> <4AE6CA4D.9010604@redhat.com> <20091027123749.GA4530@redhat.com> <20091027144209.GA13805@linuxtx.org> <4AE71264.7000102@redhat.com> <1256658393.2803.31.camel@fedora64.linuxtx.org> Message-ID: <4AE71880.7060601@redhat.com> On 10/27/2009 05:46 PM, Justin M. Forbes wrote: > On Tue, 2009-10-27 at 17:31 +0200, Izik Eidus wrote: > >> On 10/27/2009 04:42 PM, Justin M. Forbes wrote: >> >> >>> While I do understand what you are saying, I don't think it is worth making >>> a kernel change for at this point in the cycle. Because ksm itself has a >>> separate initscript, people who wish to use ksm will likely turn it on. >>> This sets the max_kernel_pages to a reasonable value. People who are not >>> interested enough to turn on the ksm service are probably not the kind of >>> people who will be checking to see how effective ksm is at all. >>> >>> >> To me it sound that users have no idea about this script, and ksm merge >> to him just the zero pages of windows 7... >> >> My feeling is that 99% of the ppl in the world that will use it, would >> just see the zero page merged and think "that is it..." >> The current behaivor in fedora 12 is misleading the user (at least it >> seems to me that it misslead that specific user) >> >> Btw we can set this value from userspace if we want "echo 0> >> /sys/kernel/mm/ksm/run" >> > Where would we put that in userspace? Another init script? I just > updated to documentation on the feature page (and hopefully in the final > release notes) to say that ksm can be enabled by running 'sudo chkconfig > ksm on' in effort to draw more attention to these scripts. > > The other option is to make the ksm init script default to on, and just > leave ksmtuned off at system start. Since ksm is on by the default > kernel, all we would be doing here is changing max_kernel_pages to a > machine specific value. The ksm init script is part of the qemu > package, so it would only be changed on systems doing virt anyway. > Sound Great to me, Can we please do that? (I am sorry I came at such late time, but I saw that ppl just run ./qemu and think ksm is working effective when they see the memory shrink (while it truely just merge the zero pages for them - due to the 2000 pages limit)) Thanks. > Justin > > From berrange at redhat.com Tue Oct 27 16:05:06 2009 From: berrange at redhat.com (Daniel P. Berrange) Date: Tue, 27 Oct 2009 16:05:06 +0000 Subject: [fedora-virt] disabling ksm by default In-Reply-To: <1256658916.25282.75.camel@blaa> References: <4AE628CA.1010807@redhat.com> <1256638369.25282.60.camel@blaa> <4AE6CA4D.9010604@redhat.com> <20091027123749.GA4530@redhat.com> <20091027144209.GA13805@linuxtx.org> <4AE71264.7000102@redhat.com> <1256658393.2803.31.camel@fedora64.linuxtx.org> <1256658916.25282.75.camel@blaa> Message-ID: <20091027160506.GF26540@redhat.com> On Tue, Oct 27, 2009 at 03:55:16PM +0000, Mark McLoughlin wrote: > On Tue, 2009-10-27 at 10:46 -0500, Justin M. Forbes wrote: > > On Tue, 2009-10-27 at 17:31 +0200, Izik Eidus wrote: > > > On 10/27/2009 04:42 PM, Justin M. Forbes wrote: > For Fedora 13, it'll be off by default in the kernel and the recommended > way of switching it on is with 'chkconfig ksm on' > > For Fedora 12, it's on by default in the kernel, 'chkconfig ksm on' just > changes max pages and the only way of disabling it is by manually > writing zero to /sys/kernel/mm/ksm/run > > IMHO, if we are to change anything for Fedora 12, we should move closer > to Fedora 13 behaviour, not further away from it - i.e. make it off by > default in the kernel, but I'm fine with delaying that until post GA > > We'll probably end up with this behaviour in F12 updates at some point > anyway when 2.6.32 is pulled in Agreed, we should change our custom patches to match upstream, so its off by default. Daniel -- |: Red Hat, Engineering, London -o- http://people.redhat.com/berrange/ :| |: http://libvirt.org -o- http://virt-manager.org -o- http://ovirt.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: GnuPG: 7D3B9505 -o- F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 :| From giallu at gmail.com Tue Oct 27 16:07:19 2009 From: giallu at gmail.com (Gianluca Sforna) Date: Tue, 27 Oct 2009 17:07:19 +0100 Subject: [fedora-virt] Porting existing virtual machines to newer Fedora In-Reply-To: <1256638129.25282.57.camel@blaa> References: <1256638129.25282.57.camel@blaa> Message-ID: On Tue, Oct 27, 2009 at 11:08 AM, Mark McLoughlin wrote: > On Mon, 2009-10-26 at 23:29 +0100, Gianluca Sforna wrote: >> I'm installing Fedora 12 (full install, no upgrade) on my laptop where >> I have some virtual machines configured via F11's virt-manager. >> >> After installation I'd like to recover from a backup the existing >> virtual machines; is there any documentation on the >> proper/recommended/safer method for doing the job? > > Not sure if we have any nice documentation on this, but it should simply > be a matter of backing up and restoring the guest's XML configuration > and disk images. Thanks, I will try ad let you know if I run into any problems. -- Gianluca Sforna http://morefedora.blogspot.com http://www.linkedin.com/in/gianlucasforna From jmforbes at linuxtx.org Tue Oct 27 16:08:08 2009 From: jmforbes at linuxtx.org (Justin M. Forbes) Date: Tue, 27 Oct 2009 11:08:08 -0500 Subject: [fedora-virt] disabling ksm by default In-Reply-To: <1256658916.25282.75.camel@blaa> References: <4AE628CA.1010807@redhat.com> <1256638369.25282.60.camel@blaa> <4AE6CA4D.9010604@redhat.com> <20091027123749.GA4530@redhat.com> <20091027144209.GA13805@linuxtx.org> <4AE71264.7000102@redhat.com> <1256658393.2803.31.camel@fedora64.linuxtx.org> <1256658916.25282.75.camel@blaa> Message-ID: <1256659688.2803.34.camel@fedora64.linuxtx.org> On Tue, 2009-10-27 at 15:55 +0000, Mark McLoughlin wrote: > On Tue, 2009-10-27 at 10:46 -0500, Justin M. Forbes wrote: > > On Tue, 2009-10-27 at 17:31 +0200, Izik Eidus wrote: > > > On 10/27/2009 04:42 PM, Justin M. Forbes wrote: > > > > > > > While I do understand what you are saying, I don't think it is worth making > > > > a kernel change for at this point in the cycle. Because ksm itself has a > > > > separate initscript, people who wish to use ksm will likely turn it on. > > > > This sets the max_kernel_pages to a reasonable value. People who are not > > > > interested enough to turn on the ksm service are probably not the kind of > > > > people who will be checking to see how effective ksm is at all. > > > > > > > To me it sound that users have no idea about this script, and ksm merge > > > to him just the zero pages of windows 7... > > > > > > My feeling is that 99% of the ppl in the world that will use it, would > > > just see the zero page merged and think "that is it..." > > > The current behaivor in fedora 12 is misleading the user (at least it > > > seems to me that it misslead that specific user) > > > > > > Btw we can set this value from userspace if we want "echo 0 > > > > /sys/kernel/mm/ksm/run" > > > > Where would we put that in userspace? Another init script? I just > > updated to documentation on the feature page (and hopefully in the final > > release notes) to say that ksm can be enabled by running 'sudo chkconfig > > ksm on' in effort to draw more attention to these scripts. > > > > The other option is to make the ksm init script default to on, and just > > leave ksmtuned off at system start. Since ksm is on by the default > > kernel, all we would be doing here is changing max_kernel_pages to a > > machine specific value. The ksm init script is part of the qemu > > package, so it would only be changed on systems doing virt anyway. > > For Fedora 13, it'll be off by default in the kernel and the recommended > way of switching it on is with 'chkconfig ksm on' > > For Fedora 12, it's on by default in the kernel, 'chkconfig ksm on' just > changes max pages and the only way of disabling it is by manually > writing zero to /sys/kernel/mm/ksm/run > > IMHO, if we are to change anything for Fedora 12, we should move closer > to Fedora 13 behaviour, not further away from it - i.e. make it off by > default in the kernel, but I'm fine with delaying that until post GA > > We'll probably end up with this behaviour in F12 updates at some point > anyway when 2.6.32 is pulled in This is exactly why I am more concerned with people using the init script than changing something else. If people use the init script, behavior doesn't change when F12 updates to 2.6.32. Whether or not we default the ksm init script to on might be a conversation worth having, but making other changes just seems a bit hackish at this point. Justin From jmforbes at linuxtx.org Tue Oct 27 16:11:54 2009 From: jmforbes at linuxtx.org (Justin M. Forbes) Date: Tue, 27 Oct 2009 11:11:54 -0500 Subject: [fedora-virt] disabling ksm by default In-Reply-To: <20091027160506.GF26540@redhat.com> References: <4AE628CA.1010807@redhat.com> <1256638369.25282.60.camel@blaa> <4AE6CA4D.9010604@redhat.com> <20091027123749.GA4530@redhat.com> <20091027144209.GA13805@linuxtx.org> <4AE71264.7000102@redhat.com> <1256658393.2803.31.camel@fedora64.linuxtx.org> <1256658916.25282.75.camel@blaa> <20091027160506.GF26540@redhat.com> Message-ID: <1256659914.2803.37.camel@fedora64.linuxtx.org> On Tue, 2009-10-27 at 16:05 +0000, Daniel P. Berrange wrote: > On Tue, Oct 27, 2009 at 03:55:16PM +0000, Mark McLoughlin wrote: > > On Tue, 2009-10-27 at 10:46 -0500, Justin M. Forbes wrote: > > > On Tue, 2009-10-27 at 17:31 +0200, Izik Eidus wrote: > > > > On 10/27/2009 04:42 PM, Justin M. Forbes wrote: > > For Fedora 13, it'll be off by default in the kernel and the recommended > > way of switching it on is with 'chkconfig ksm on' > > > > For Fedora 12, it's on by default in the kernel, 'chkconfig ksm on' just > > changes max pages and the only way of disabling it is by manually > > writing zero to /sys/kernel/mm/ksm/run > > > > IMHO, if we are to change anything for Fedora 12, we should move closer > > to Fedora 13 behaviour, not further away from it - i.e. make it off by > > default in the kernel, but I'm fine with delaying that until post GA > > > > We'll probably end up with this behaviour in F12 updates at some point > > anyway when 2.6.32 is pulled in > > Agreed, we should change our custom patches to match upstream, so its > off by default. We have done this twice now, initially ksm defaulted to off, and to merge with what was to become 2.6.32 we ended up with the default on and the current page limits. Of course upstream has changed again. Behavior doesn't change with the init script. People really don't know how to check if ksm is on or off without checking the documentation. As long as our documentation points to the initscript instead of modifying sysfs we get the desired behavior. Justin From overholt at redhat.com Thu Oct 29 17:25:18 2009 From: overholt at redhat.com (Andrew Overholt) Date: Thu, 29 Oct 2009 13:25:18 -0400 Subject: [fedora-virt] F-12 host, can't acquire DHCP lease in guest Message-ID: <20091029172518.GE2214@redhat.com> Hi, I installed rawhide/F-12 x86_64 (network using NetworkManager) and installed and started libvirtd. I then created both F-12 i686 and F-11 i686 guests which went fine. However, neither guest is able to acquire a DHCP lease. Is there some sort of configuration that I didn't do? I don't recall having to do anything special with F-11 as a host. Relevant information below. I can file a bug if that's preferred. Thanks, Andrew ------------------------------------------------------------------------ # virsh net-list --all Name State Autostart ----------------------------------------- default active yes # brctl show bridge name bridge id STP enabled interfaces virbr0 8000.6e8e6e10bdb6 yes vnet0 vnet1 # sysctl net.bridge.bridge-nf-call-iptables net.bridge.bridge-nf-call-iptables = 1 # iptables -L -v -n Chain INPUT (policy ACCEPT 0 packets, 0 bytes) pkts bytes target prot opt in out source destination 2129K 2744M ACCEPT all -- * * 0.0.0.0/0 0.0.0.0/0 state RELATED,ESTABLISHED 3 1070 ACCEPT icmp -- * * 0.0.0.0/0 0.0.0.0/0 7 420 ACCEPT all -- lo * 0.0.0.0/0 0.0.0.0/0 0 0 ACCEPT tcp -- * * 0.0.0.0/0 0.0.0.0/0 state NEW tcp dpt:22 2669 335K REJECT all -- * * 0.0.0.0/0 0.0.0.0/0 reject-with icmp-host-prohibited Chain FORWARD (policy ACCEPT 0 packets, 0 bytes) pkts bytes target prot opt in out source destination 9 2952 ACCEPT all -- * * 0.0.0.0/0 0.0.0.0/0 PHYSDEV match --physdev-is-bridged 0 0 REJECT all -- * * 0.0.0.0/0 0.0.0.0/0 reject-with icmp-host-prohibited Chain OUTPUT (policy ACCEPT 1439K packets, 1128M bytes) pkts bytes target prot opt in out source destination # ps -ef | grep dnsmasq nobody 3148 1 0 09:43 ? 00:00:00 /usr/sbin/dnsmasq --strict-order --bind-interfaces --pid-file=/var/run/libvirt/network/default.pid --conf-file= --listen-address 192.168.122.1 --except-interface lo --dhcp-range 192.168.122.2,192.168.122.254 root 20983 20857 0 13:23 pts/5 00:00:00 grep dnsmasq # ifconfig -a eth0 Link encap:Ethernet HWaddr 00:23:54:9E:57:84 inet addr:10.15.16.87 Bcast:10.15.16.255 Mask:255.255.255.0 inet6 addr: fe80::223:54ff:fe9e:5784/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:2414348 errors:0 dropped:0 overruns:0 frame:0 TX packets:2144264 errors:0 dropped:0 overruns:0 carrier:2 collisions:0 txqueuelen:1000 RX bytes:2888713375 (2.6 GiB) TX bytes:1065980118 (1016.5 MiB) Interrupt:29 lo Link encap:Local Loopback inet addr:127.0.0.1 Mask:255.0.0.0 inet6 addr: ::1/128 Scope:Host UP LOOPBACK RUNNING MTU:16436 Metric:1 RX packets:58553 errors:0 dropped:0 overruns:0 frame:0 TX packets:58553 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:256208705 (244.3 MiB) TX bytes:256208705 (244.3 MiB) virbr0 Link encap:Ethernet HWaddr 6E:8E:6E:10:BD:B6 inet addr:192.168.122.1 Bcast:192.168.122.255 Mask:255.255.255.0 UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:85 errors:0 dropped:0 overruns:0 frame:0 TX packets:30 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:19960 (19.4 KiB) TX bytes:6015 (5.8 KiB) vnet0 Link encap:Ethernet HWaddr 6E:8E:6E:10:BD:B6 inet6 addr: fe80::6c8e:6eff:fe10:bdb6/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:14 errors:0 dropped:0 overruns:0 frame:0 TX packets:3410 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:500 RX bytes:3204 (3.1 KiB) TX bytes:180086 (175.8 KiB) vnet1 Link encap:Ethernet HWaddr 92:AE:24:5C:77:7E inet6 addr: fe80::90ae:24ff:fe5c:777e/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:15 errors:0 dropped:0 overruns:0 frame:0 TX packets:3211 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:500 RX bytes:3546 (3.4 KiB) TX bytes:167128 (163.2 KiB) # cat /proc/sys/net/ipv4/ip_forward 1 I've reloaded the libvirtd service, restarted it, rebooted the guests, etc. From markmc at redhat.com Thu Oct 29 17:45:56 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Thu, 29 Oct 2009 17:45:56 +0000 Subject: [fedora-virt] F-12 host, can't acquire DHCP lease in guest In-Reply-To: <20091029172518.GE2214@redhat.com> References: <20091029172518.GE2214@redhat.com> Message-ID: <1256838356.10825.107.camel@blaa> On Thu, 2009-10-29 at 13:25 -0400, Andrew Overholt wrote: > Hi, > > I installed rawhide/F-12 x86_64 (network using NetworkManager) and installed > and started libvirtd. I then created both F-12 i686 and F-11 i686 guests which > went fine. However, neither guest is able to acquire a DHCP lease. > > Is there some sort of configuration that I didn't do? I don't recall having to > do anything special with F-11 as a host. Relevant information below. > > I can file a bug if that's preferred. > > Thanks, > > Andrew > > ------------------------------------------------------------------------ > > # virsh net-list --all > Name State Autostart > ----------------------------------------- > default active yes > > # brctl show > bridge name bridge id STP enabled interfaces > virbr0 8000.6e8e6e10bdb6 yes vnet0 > vnet1 > > # sysctl net.bridge.bridge-nf-call-iptables > net.bridge.bridge-nf-call-iptables = 1 This is supposed to be zero by default in F-12: https://bugzilla.redhat.com/512206 > # iptables -L -v -n > Chain INPUT (policy ACCEPT 0 packets, 0 bytes) > pkts bytes target prot opt in out source destination > 2129K 2744M ACCEPT all -- * * 0.0.0.0/0 0.0.0.0/0 state RELATED,ESTABLISHED > 3 1070 ACCEPT icmp -- * * 0.0.0.0/0 0.0.0.0/0 > 7 420 ACCEPT all -- lo * 0.0.0.0/0 0.0.0.0/0 > 0 0 ACCEPT tcp -- * * 0.0.0.0/0 0.0.0.0/0 state NEW tcp dpt:22 > 2669 335K REJECT all -- * * 0.0.0.0/0 0.0.0.0/0 reject-with icmp-host-prohibited Looks like iptables is missing the rules libvirt starts; a 'service libvirtd reload' should load them again, but you say you've tried that? Does 'virsh net-destroy default' and 'virsh net-start default' followed by a restart of the guests help? Background on libvirt/iptables integration here: https://bugzilla.redhat.com/227011 Cheers, Mark. From overholt at redhat.com Thu Oct 29 18:19:53 2009 From: overholt at redhat.com (Andrew Overholt) Date: Thu, 29 Oct 2009 14:19:53 -0400 Subject: [fedora-virt] F-12 host, can't acquire DHCP lease in guest In-Reply-To: <1256838356.10825.107.camel@blaa> References: <20091029172518.GE2214@redhat.com> <1256838356.10825.107.camel@blaa> Message-ID: <20091029181952.GG2214@redhat.com> Hi, * Mark McLoughlin [2009-10-29 13:47]: > On Thu, 2009-10-29 at 13:25 -0400, Andrew Overholt wrote: > > # sysctl net.bridge.bridge-nf-call-iptables > > net.bridge.bridge-nf-call-iptables = 1 > > This is supposed to be zero by default in F-12: > > https://bugzilla.redhat.com/512206 I've commented. > > # iptables -L -v -n > > Chain INPUT (policy ACCEPT 0 packets, 0 bytes) > > pkts bytes target prot opt in out source destination > > 2129K 2744M ACCEPT all -- * * 0.0.0.0/0 0.0.0.0/0 state RELATED,ESTABLISHED > > 3 1070 ACCEPT icmp -- * * 0.0.0.0/0 0.0.0.0/0 > > 7 420 ACCEPT all -- lo * 0.0.0.0/0 0.0.0.0/0 > > 0 0 ACCEPT tcp -- * * 0.0.0.0/0 0.0.0.0/0 state NEW tcp dpt:22 > > 2669 335K REJECT all -- * * 0.0.0.0/0 0.0.0.0/0 reject-with icmp-host-prohibited > > Looks like iptables is missing the rules libvirt starts; a 'service > libvirtd reload' should load them again, but you say you've tried that? Yeah, that didn't seem to do anything. I re-tried it and re-started my guests but nothing changed. > Does 'virsh net-destroy default' and 'virsh net-start default' followed > by a restart of the guests help? Yes, that fixes it. Is there a bug I should file? TVM, Andrew From rjones at redhat.com Thu Oct 29 18:59:04 2009 From: rjones at redhat.com (Richard W.M. Jones) Date: Thu, 29 Oct 2009 18:59:04 +0000 Subject: [fedora-virt] Got Windows guests? Message-ID: <20091029185904.GA8383@amd.home.annexia.org> Hello fellow Fedora, libvirt and libguestfs users, If you have any Windows guests, then you can help Fedora to support Windows guests better by spending a few minutes testing the Windows Registry feature we just added to libguestfs 1.0.75. You will need: - A Windows NT/200x/XP/Vista/7/... guest - Fedora 12 or Fedora Rawhide host - libguestfs-tools >= 1.0.75 (from updates or http://koji.fedoraproject.org/koji/packageinfo?packageID=8391 ) - a few minutes of your time The tests: (1) Run the virt-win-reg commands shown in the following web page, where "MyWinGuest" should be replaced with the name of the Windows guest as known to libvirt: http://libguestfs.org/virt-win-reg.1.html#examples Do the commands run without any errors? Does the output look sensible? If you have several Windows guests, please try as many different sorts as possible! (2) Download the registry binary files and try to convert them to XML: guestfish -i MyWinGuest --ro <<'EOF' download win:\windows\system32\config\software software download win:\windows\system32\config\system system download win:\windows\system32\config\sam sam download win:\windows\system32\config\security security EOF hivexml software > software.xml hivexml system > system.xml hivexml sam > sam.xml hivexml security > security.xml Do those commands run without error? If there's an error, try adding the hivexml -k option. Does the XML look complete? (Try running the XML through tidy -xml -indent -quiet < foo.xml | less ) I hope you don't find any bugs, but if you do: Send a reply to this message, or report a bug in Bugzilla: https://bugzilla.redhat.com/enter_bug.cgi?component=libguestfs&product=Virtualization+Tools It's useful to include the following details: HIVEX_DEBUG=1 hivexml regfile 2>&1 > log.out The Registry file itself that is failing (but note that Registry files can contain sensitive data). It's also useful to have positive feedback ("it worked!"). Thanks for any testing you can give, and if you have any other suggestions for handling Windows guests from Fedora, please let me know. Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones virt-df lists disk usage of guests without needing to install any software inside the virtual machine. Supports Linux and Windows. http://et.redhat.com/~rjones/virt-df/ From dennisml at conversis.de Thu Oct 29 19:35:58 2009 From: dennisml at conversis.de (Dennis J.) Date: Thu, 29 Oct 2009 20:35:58 +0100 Subject: [fedora-virt] Trying to move VM efficiently, grub-install, guestfish Message-ID: <4AE9EE9E.20605@conversis.de> Hi, What I'm currently experimenting with is how to move a VM between two hosts efficiently. Specifically we are talking Linux guest with virtual disk sizes of 150G of which only about 10-20G are used. Doing a plain "dd" takes a lot of time and is kind of wasteful given that 90% of the logical volume used doesn't contain relevant data. So right now I'm trying to re-create the lvm/filesystem structure from scratch on the target system and then copying the data using tar from the source system. This looks good so far but I cannot start the guest on the target because the boot-sector is missing. I tried using guestfish as described here: http://rwmj.wordpress.com/2009/05/11/rescuing-a-failed-vm-install-with-guestfish/ but when I issue the "grub-install / /dev/sda" all I get is: "libguestfs: error: grub_install: grub-install: /dev/sda does not have any corresponding BIOS drive." I also tried copying the MBR using "dd bs=448 count=1 ..." but that only results in the output of "GRUB" on the target system and then nothing happens. What is the best way to copy/initialize the MBR so the target system boots again? Another method I intend to test is to copy the entire volume with "dd" while the VM is still running then shutting the VM down and using "rsync" to do the final fixup on the logical volumes. That of course assumes that rsync actually works on device nodes which I'm not so sure about. Has anyone experience with this kind of thing? (Just to be clear I'm not talking about cloning the VM requiring making changes to the files in the VM but just simply moving the VM to another host with as little down-time as possible.) Regards, Dennis From orion at cora.nwra.com Thu Oct 29 20:19:42 2009 From: orion at cora.nwra.com (Orion Poplawski) Date: Thu, 29 Oct 2009 14:19:42 -0600 Subject: [fedora-virt] Strange pygrub issue Message-ID: <4AE9F8DE.1010307@cora.nwra.com> I've got a strange problem with one of my Fedora 11 guests on CentOS 5.4 host. pygrub does not report the current grub menu, but an old (previous) one. So the system boots to an old kernel, despite everything seeming to be up to date in the guest. Any ideas? -- Orion Poplawski Technical Manager 303-415-9701 x222 NWRA/CoRA Division FAX: 303-415-9702 3380 Mitchell Lane orion at cora.nwra.com Boulder, CO 80301 http://www.cora.nwra.com From rjones at redhat.com Thu Oct 29 20:28:37 2009 From: rjones at redhat.com (Richard W.M. Jones) Date: Thu, 29 Oct 2009 20:28:37 +0000 Subject: [fedora-virt] Trying to move VM efficiently, grub-install, guestfish In-Reply-To: <4AE9EE9E.20605@conversis.de> References: <4AE9EE9E.20605@conversis.de> Message-ID: <20091029202837.GB8383@amd.home.annexia.org> On Thu, Oct 29, 2009 at 08:35:58PM +0100, Dennis J. wrote: > http://rwmj.wordpress.com/2009/05/11/rescuing-a-failed-vm-install-with-guestfish/ > but when I issue the "grub-install / /dev/sda" all I get is: > "libguestfs: error: grub_install: grub-install: /dev/sda does not have > any corresponding BIOS drive." I read the rest of your comment with awe. All I can say is good luck with your experiments. Just about the particular error above .. Is /dev/sda the real name of the drive? (Try the list-devices command to list the real device names). Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones virt-top is 'top' for virtual machines. Tiny program with many powerful monitoring features, net stats, disk stats, logging, etc. http://et.redhat.com/~rjones/virt-top From dennisml at conversis.de Thu Oct 29 22:06:41 2009 From: dennisml at conversis.de (Dennis J.) Date: Thu, 29 Oct 2009 23:06:41 +0100 Subject: [fedora-virt] Trying to move VM efficiently, grub-install, guestfish In-Reply-To: <20091029202837.GB8383@amd.home.annexia.org> References: <4AE9EE9E.20605@conversis.de> <20091029202837.GB8383@amd.home.annexia.org> Message-ID: <4AEA11F1.2090303@conversis.de> On 10/29/2009 09:28 PM, Richard W.M. Jones wrote: > On Thu, Oct 29, 2009 at 08:35:58PM +0100, Dennis J. wrote: >> http://rwmj.wordpress.com/2009/05/11/rescuing-a-failed-vm-install-with-guestfish/ >> but when I issue the "grub-install / /dev/sda" all I get is: >> "libguestfs: error: grub_install: grub-install: /dev/sda does not have >> any corresponding BIOS drive." > > I read the rest of your comment with awe. All I can say is good luck > with your experiments. > > Just about the particular error above .. Is /dev/sda the real name of > the drive? (Try the list-devices command to list the real device > names). Success! The problem was that the guest is usually using /dev/vda but I couldn't get guestfish to start with: [root at nexus ~]# guestfish -a /var/lib/libvirt/images/test2.img -m /dev/vg_test/lv_root -m /dev/vda1:/boot libguestfs: error: mount: mount_stub: /dev/vda1: No such file or directory Using /dev/sda1 instead lets guestfish start and /boot is mounted properly but grub-install didn't work. What did the trick was changing "/dev/vda" to "/dev/sda" in /boot/grub/device.map followed by a grub-install which now works and then changing the device.map back so I don't get problems when booting up the guest for real. After that the "new" guest boots up perfectly. I'm not very familiar with guestfish (yet) so I'm not sure if the /dev/vda1 thing is a bug or simply not supposed to work. The next step would be to automate the whole thing with a script even if it's only useful in some cases (where the guest only really uses a small part of the image). Regards, Dennis From rjones at redhat.com Thu Oct 29 22:30:24 2009 From: rjones at redhat.com (Richard W.M. Jones) Date: Thu, 29 Oct 2009 22:30:24 +0000 Subject: [fedora-virt] Trying to move VM efficiently, grub-install, guestfish In-Reply-To: <4AEA11F1.2090303@conversis.de> References: <4AE9EE9E.20605@conversis.de> <20091029202837.GB8383@amd.home.annexia.org> <4AEA11F1.2090303@conversis.de> Message-ID: <20091029223024.GA9681@amd.home.annexia.org> On Thu, Oct 29, 2009 at 11:06:41PM +0100, Dennis J. wrote: > I'm not very familiar with guestfish (yet) so I'm not sure if the > /dev/vda1 thing is a bug or simply not supposed to work. Reliable device naming is always a problem with Linux. This explains the situation vis-a-vis libguestfs clients: http://libguestfs.org/guestfs.3.html#block_device_naming Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones Read my programming blog: http://rwmj.wordpress.com Fedora now supports 80 OCaml packages (the OPEN alternative to F#) http://cocan.org/getting_started_with_ocaml_on_red_hat_and_fedora From clalance at redhat.com Fri Oct 30 08:13:32 2009 From: clalance at redhat.com (Chris Lalancette) Date: Fri, 30 Oct 2009 09:13:32 +0100 Subject: [fedora-virt] Strange pygrub issue In-Reply-To: <4AE9F8DE.1010307@cora.nwra.com> References: <4AE9F8DE.1010307@cora.nwra.com> Message-ID: <4AEAA02C.4000105@redhat.com> Orion Poplawski wrote: > I've got a strange problem with one of my Fedora 11 guests on CentOS 5.4 > host. pygrub does not report the current grub menu, but an old > (previous) one. So the system boots to an old kernel, despite > everything seeming to be up to date in the guest. > > Any ideas? It's most likely: https://bugzilla.redhat.com/show_bug.cgi?id=466681 The workaround is to: # echo 3 > /proc/sys/vm/drop_caches -- Chris Lalancette From rjones at redhat.com Fri Oct 30 10:00:28 2009 From: rjones at redhat.com (Richard W.M. Jones) Date: Fri, 30 Oct 2009 10:00:28 +0000 Subject: [fedora-virt] disabling ksm by default In-Reply-To: <1256638369.25282.60.camel@blaa> References: <4AE628CA.1010807@redhat.com> <1256638369.25282.60.camel@blaa> Message-ID: <20091030100028.GA12068@amd.home.annexia.org> On Tue, Oct 27, 2009 at 10:12:49AM +0000, Mark McLoughlin wrote: > ksm is disabled by default currently [...] > # unless KSM_MAX_KERNEL_PAGES is set, let ksm munch up to half of total memory. Mark, can you summarise why we wouldn't want KSM to be enabled by default, and why we need to limit it to half of total memory? Assume the reader knows nothing about KSM, which is approximately correct :-) Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones Read my programming blog: http://rwmj.wordpress.com Fedora now supports 80 OCaml packages (the OPEN alternative to F#) http://cocan.org/getting_started_with_ocaml_on_red_hat_and_fedora From jmforbes at linuxtx.org Fri Oct 30 15:04:20 2009 From: jmforbes at linuxtx.org (Justin M. Forbes) Date: Fri, 30 Oct 2009 10:04:20 -0500 Subject: [fedora-virt] disabling ksm by default In-Reply-To: <20091030100028.GA12068@amd.home.annexia.org> References: <4AE628CA.1010807@redhat.com> <1256638369.25282.60.camel@blaa> <20091030100028.GA12068@amd.home.annexia.org> Message-ID: <20091030150420.GA22371@linuxtx.org> On Fri, Oct 30, 2009 at 10:00:28AM +0000, Richard W.M. Jones wrote: > On Tue, Oct 27, 2009 at 10:12:49AM +0000, Mark McLoughlin wrote: > > ksm is disabled by default currently > [...] > > # unless KSM_MAX_KERNEL_PAGES is set, let ksm munch up to half of total memory. > > Mark, can you summarise why we wouldn't want KSM to be enabled by > default, and why we need to limit it to half of total memory? > The limit to half of total memory is because ksm pages are unswappable at this time. To be fixed in a future kernel. As for enabling by default, the more I think about it, the more this makes sense. The ksm initscript is shipped with qemu, so anyone with it installed should be interested in virt, and the value of max_kernel_pages doesn't matter if you are not running anything which marks memory mergeable, which only kvm does at the moment. Justin From markmc at redhat.com Fri Oct 30 15:42:32 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Fri, 30 Oct 2009 15:42:32 +0000 Subject: [fedora-virt] virt-preview updates In-Reply-To: <1250703309.17874.46.camel@blaa> References: <1244034684.5001.134.camel@blaa> <1244046483.5001.180.camel@blaa> <1244197459.27876.0.camel@blaa> <1246303956.11688.148.camel@blaa> <1246628572.13604.2.camel@blaa> <1246636154.13604.8.camel@blaa> <1247763853.3038.105.camel@blaa> <1248769817.3089.20.camel@blaa> <1248776732.3089.45.camel@blaa> <1248859649.3021.28.camel@blaa> <20090729093637.GC10723@salstar.sk> <1248868438.3021.31.camel@blaa> <1248873640.3021.34.camel@blaa> <1248879462.3021.39.camel@blaa> <1248882336.3021.40.camel@blaa> <1248974623.3275.4.camel@blaa> <1249057022.3320.35.camel@blaa> <1249400197.3212.73.camel@blaa> <1249573056.3721.62.camel@blaa> <1249639097.3260.5.camel@blaa> <1249900612.8784.79.camel@blaa> <1250700893.17874.44.camel@blaa> <1250703309.17874.46.camel@blaa> Message-ID: <1256917352.6899.129.camel@blaa> The latest updates available from http://markmc.fedorapeople.org/virt-preview == qemu == * Thu Oct 29 2009 Mark McLoughlin - 2:0.11.0-9 - Fix dropped packets with non-virtio NICs (#531419) * Wed Oct 21 2009 Glauber Costa - 2:0.11.0-8 - Properly save kvm time registers (#524229) == libvirt == * Thu Oct 29 2009 Mark McLoughlin - 0.7.1-15 - Avoid compressing small log files (#531030) * Thu Oct 29 2009 Mark McLoughlin - 0.7.1-14 - Make libvirt-devel require libvirt-client, not libvirt - Fix xen driver recounting (#531429) - Fix crash on virsh error (#531429) - Fix segfault where XML parsing fails in qemu disk hotplug - Fix segfault where interface target device name is ommitted (#523418) Cheers, Mark. == About The Virtualization Preview Repository == The virt-preview repository offers Fedora 11 users the opportunity to help out with testing the rawhide virtualization packages, without having to run all of rawhide. Bugs should be reported to http://bugzilla.redhat.com/ using 'rawhide' as the distribution version. Note clearly in the bug report that you are running these packages on Fedora 11. From markmc at redhat.com Fri Oct 30 16:44:26 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Fri, 30 Oct 2009 16:44:26 +0000 Subject: [fedora-virt] Fedora virt status Message-ID: <1256921066.6899.148.camel@blaa> Fedora 12 ========= Fedora 12 Beta slipped by a week, but was released on Oct 20: https://fedoraproject.org/wiki/F12_Beta_Announcement We're now in the final countdown to F12 GA. The first release candidate will be composed next week and, all going well, we should have a release on November 17. Thankfully, the virt blocker list is now clear, but if you're looking to help with making Fedora 12 even better, there's no better place to start than the F12 target tracker bug: https://bugzilla.redhat.com/showdependencytree.cgi?id=F12VirtTarget&hide_resolved=1 There's over 100 bugs there that need your help! Witty Tagline ============= Dan Berrange has done a great job of coming up with a "witty tagline" for each of the virtualization feature sets in previous Fedora releases: https://fedoraproject.org/wiki/Virtualization/History Fedora 11: The walled garden Fedora 10: Management at a distance Fedora 9: Farewell to old friends Fedora 8: Protection from the bad guys Fedora 7: The new kid on the block Fedora Core 6: Virtualization grows up Fedora Core 5: The future is now Fedora Core 4: Glimpse of the future We need a tagline which describes our eclectic mix of features in Fedora 12. Any ideas? Fedora Weekly News ================== Dale Bewley has contributed another bunch of virtualization sections to Fedora Weekly News: https://fedoraproject.org/wiki/FWN/Issue196#Virtualization https://fedoraproject.org/wiki/FWN/Issue197#Virtualization https://fedoraproject.org/wiki/FWN/Issue199#Virtualization ABRT and qemu-kvm ================= It's quite common that we have to ask bug reports to collect a stack trace for a qemu-kvm segfault. Is ABRT the future? https://bugzilla.redhat.com/show_bug.cgi?id=528754#c1 Try this: $> yum install --enablerepo=rawhide-debuginfo qemu-debuginfo $> yum install -y abrt abrt-cli abrt-addon-ccpp abrt-plugin-logger $> service abrtd restart then when qemu-kvm crashes, you should be able to do e.g.: $> abrt-cli --get-list $> echo n | abrt-cli --report $uuid > t.log Anyone got ideas for a better way to use ABRT to catch virt related bugs? We could do with adding instructions to: http://fedoraproject.org/wiki/How_to_debug_Virtualization_problems Got Windows Guests? =================== Rich Jones has added yet another weird and fantastical feature to libguestfs ... and he needs your help to check that it works: http://www.redhat.com/archives/fedora-virt/2009-October/msg00128.html If you have any Windows guests, then you can help Fedora to support Windows guests better by spending a few minutes testing the Windows Registry feature we just added to libguestfs 1.0.75. Basically, the new virt-win-reg tool can be used to display entries from the registry in a Windows guest. How cool is that? Bugzilla ======== DOOM-O-METER: 198 bugs open four weeks ago, only 186 now! Ongoing Bugs ============ == kernel == https://bugzilla.redhat.com/515741 2.6.30 kernel stopped supporting xattrs on hugetlbfs John Cooper is working on getting a patch together for F-12. https://bugzilla.redhat.com/524808 swiotlb should be enabled when VT-d setup fails dwmw2 points out that if VT-d setup fails (because e.g. a bad BIOS is detected), then we should make sure that swiotlb is still enabled. Chris Wright has been working on the fix. All are agreed that this shouldn't be considered a release blocker. https://bugzilla.redhat.com/530389 allow userspace to adjust kvmclock offset Glauber has patches queued upstream to add new kvmclock ioctls and wants them backported to Fedora 12. https://bugzilla.redhat.com/531270 TG3, kvm, ipv6 & tso data corruption bug Rik van Riel discovered this TSO6 related issue with the tg3 driver which is causing traffic from his KVM guest to be corrupted. https://bugzilla.redhat.com/528245 starting a KVM guest before wireless NIC is brought up results in "nf_conntrack: table full, dropping packet." Paul Lambert reports that his machine locks up on boot if he sets his KVM guest to autostart. It looks like it may be related to the wireless connection not being available at that point, but we're still pretty far from narrowing the issue down. == qemu == https://bugzilla.redhat.com/528754 qemu-kvm segfault caused by -soundhw es1370 Gene Czarcinski reported this qemu segfault and managed to obtain a stack trace for it. https://bugzilla.redhat.com/528134 qemu should set O_CLOEXEC on all file descriptors glibc is now forking a ptchown program under certain circumstances, and SELinux issues showed that we are leaking file descriptors into forked processes. Uli claims we should set O_CLOEXEC but Dan, Markus and Kevin disagree. https://bugzilla.redhat.com/472236 KVM guests cannot boot PXE "local" This bug was originally reported against Fedora 10, but apparently is still a problem in Fedora 12 with gPXE. https://bugzilla.redhat.com/528764 qemu's SDL backend doesn't release the console framebuffer on exit Matt Booth discovered this annoyance with qemu's SDL backend. == libvirt == https://bugzilla.redhat.com/527736 libvirt can't create storage pool volumes on a FAT32 hard disk - 'cannot set file owner ... Operation not permitted' Reported by Richard Hughes, looks like libvirt's fs storage backend just isn't handling this case correctly. https://bugzilla.redhat.com/529363 virsh restore fails - failing to re-label the save file ? sVirt related virsh restore failure. == virtinst == https://bugzilla.redhat.com/528947 virtinst: Creating LVM volume on QEMU during VM creation fails A virtinst traceback during cloning, fixed by Cole upstream but needing backporting. == virt-manager == https://bugzilla.redhat.com/530210 virt-manager setfacl() fails on filesystems not mounted with the 'acl' option Our solution for "qemu can't access my homedir" was to use setfacl to give it access. We're scuppered in the case where users have created the filesystem themselves without the 'acl' option listed in the defaults. We need a better error message for this case. https://bugzilla.redhat.com/530948 virt-manager: switching VM to fullscreen pins the window to 1st virtual desktop Stefan Assman is seeing very strange behaviour when trying to use virt-manager in fullscreen mode. https://bugzilla.redhat.com/526952 virt-manager locks up frequently using remote connection Problem reported by both Matt Booth and Chris Wright. https://bugzilla.redhat.com/527714 Entering a non-numeric capacity creates a volume with 10000 MB Another minor virt-manager quirk. https://bugzilla.redhat.com/530448 RFE: virt-manager should provide a means to refresh storage pools virsh has a 'refresh-pool' operation, but virt-manager doesn't expose it. https://bugzilla.redhat.com/529956 16x16 icon in Virtual Machine Details window https://bugzilla.redhat.com/529954 virt-manager dialog buttons do not follow the HIG https://bugzilla.redhat.com/529955 virt-manager host details window does not honor button_have_icons setting https://bugzilla.redhat.com/529657 virt-manager accidentally showing tabs in VM details https://bugzilla.redhat.com/530096 virt-manager host details window does not need a menu Another bunch of reports of virt-manager UI warts from Michael Monreal. https://bugzilla.redhat.com/529217 virt-manager storage pool window appears behind dialog This annoyance in Fedora 11 reported by James Laska is fixed upstream, but needs backporting. https://bugzilla.redhat.com/530178 virt-manager right-click menu for connections should have a delete item Yet another UI quirk, this time reported by Gianluca Cecchi. https://bugzilla.redhat.com/529842 virt-manager: Storage browser should change window title when choosing ISO "Choose Volume" was causing some confusion which could be avoided if it was changed to "Choose ISO". == xen == https://bugzilla.redhat.com/523941 kernel 2.6.31-1[24].fc12 doesn't boot in xen PV guest on 64b host Andrew Jones is busy debugging this and has successfully bisected the problem down to a specific upstream commit. Some debate in the bug as to whether we'd actually slip the release for it. In the end, it turned out to be fixed by 2.6.31.5 for unknown reasons. https://bugzilla.redhat.com/531311 2.6.32 boot fails as Xen PV guest with stackprotector Andrew also uncovered this issue with the Fedora 13 kernel. https://bugzilla.redhat.com/526627 F11 Xen DomU unstable (2.6.30) Reporters seeing consistent Fedora 11 Xen guest lockups, seemingly dependant on the clock source. https://bugzilla.redhat.com/521800 kernel backtrace: possible recursive locking detected on Xen domU A strange recursive locking warning under Xen which apparently doesn't make much sense. Resolved Bugs ============= == misc == https://bugzilla.redhat.com/530015 Add libvirt to Virtualization group Somehow when we split libvirt-client from libvirt, we failed to make sure that libvirt was still being pulled in by the Virtualization package group. https://bugzilla.redhat.com/523914 Mouse does not move in PV Xen guest under RHEL-5.4 Wrangling of this complex blocker bug continued. Peter Hutterer (the evdev maintainer) came up with a way to allow evdev to be configured so it works like before. The fixes are now in rawhide. == kernel == https://bugzilla.redhat.com/524229 Local migration of kvm guest fails in Fedora12 Alpha Marcelo thought this was a pvmmu issue fixed in 2.6.31, but apparently not. After some more testing Marcelo now suspects it's a kvmclock related issue fixed by Glauber upstream. This is now fixed in rawhide. https://bugzilla.redhat.com/526869 Add virtio_blk support cache flush (VIRTIO_BLK_F_FLUSH) Justin has pulled this 2.6.32 patch into Fedora 12. https://bugzilla.redhat.com/525890 Backport virtio patches for optimised virtio-net We've decided to wait until 2.6.32 for these patches. == qemu == https://bugzilla.redhat.com/531419 qemu issue with non-virtio NICs receiving heavy traffic volumes Reported by Scott Tsai on the qemu-devel upstream list, patch was pulled into Fedora 12. https://bugzilla.redhat.com/524734 KVM guest ext3 errors at shutdown when using virtio and a qcow2 backing file Kevin debugged this, posted a patch upstream suitable for stable and we pulled it into F-12. https://bugzilla.redhat.com/528901 qemu-kvm : msrs[] array in kvm_arch_save_regs() too small / may cause stack corruption Ulrich Obergfell provided this excellent analysis of a savevm segfault in Fedora 11. Eduardo sent a fix upstream. https://bugzilla.redhat.com/503156 qemu VNC :: xterm inside VM shows garbled text The version of qemu in Fedora 11 appears to have a broken VNC CopyRect encoding implementation and this has been seriously annoying some users. In the absence of a confirmed fix, we've just disabled the encoding for now. https://bugzilla.redhat.com/527653 /etc/init.d/ksm: line 69: [: -eq: unary operator expected Minor bug with the ksm initscript now fixed in F-12. https://bugzilla.redhat.com/527087 qemu-common postinstall scriptlet uses getent before it's installed Bruce Jerrick points out this packaging error. Fixed in F-12 now. https://bugzilla.redhat.com/526381 qcow2 performance bad under i/o load CLOSED NOTABUG since these were compressed qcow2 images. https://bugzilla.redhat.com/530059 VDE support in qemu? CLOSED CANTFIX until someone packages VDE. https://bugzilla.redhat.com/530136 guestfwd option doesn't allow supplementary chardev options like server,nowait Rich Jones points out a problem with the -net user,guestfwd= option, but since libguestfs doesn't need it anymore and the new -chardev option upstream solves the problem, this is now marked CLOSED UPSTREAM. https://bugzilla.redhat.com/528678 qemu-img does not correctly roundtrip data for VPC image format This minor issue was caught by the libvirt-TCK test suite. Kevin reports that this is NOTABUG, but rather a limitation with the vpc file format. https://bugzilla.redhat.com/526776 Add rtl8209 to gpxe-roms-qemu https://bugzilla.redhat.com/526781 Add rtl8209 to etherboot-zroms-kvm https://bugzilla.redhat.com/526243 Guest PXE booting doesn't work when using ne2k_pci NIC model https://bugzilla.redhat.com/526777 Guest PXE booting doesn't work when using ne2k_pci NIC model Fixed this in F-11 and F-12. == libvirt == https://bugzilla.redhat.com/523158 virsh restore causes libvirtd segfault virExec() Dan Berrange identified the recent change that caused this regression. Fixed now in Fedora 12. https://bugzilla.redhat.com/528162 Fix memory leak in libvirtd in F12 An upstream fix which needed backporting to F12. https://bugzilla.redhat.com/526769 libvirt logrotate settings should roll over weekly instead of daily https://bugzilla.redhat.com/531030 log file sized increased by "compression" Two requests to fix logrotate's handling of libvirt's qemu log files. Fixed for F-12. https://bugzilla.redhat.com/517157 libvirt fails to start guest on NFS even when sebool virt_use_nfs is on We never got around to backporting the fix for this to F-11 and other users had been reporting it. Now fixed in a F-11 update. https://bugzilla.redhat.com/524732 org.libvirt.manage policy kit denial This was due to there being no PolicyKit authentication agent running in the session. == virt-manager == https://bugzilla.redhat.com/461962 virt-manager should support logical pool enumeration Creating a lvm volume group using virt-manager is confusing, but Cole has fixed that upstream now. https://bugzilla.redhat.com/440591 virt-manager: Add 'double click to connect' hint to connection row tooltip Another UI annoyance fixed by Cole upstream. https://bugzilla.redhat.com/526111 virt-manager add storage volume allows an empty name field https://bugzilla.redhat.com/526077 virt-manager toolbar buttons should have tooltips https://bugzilla.redhat.com/524083 virt-manager storage "Allocation" field can be set higher than "Max Capacity" Fixed by virt-manager-0.8.0-7.fc12 == xen == https://bugzilla.redhat.com/521806 Unowned directory /usr/lib64/fs https://bugzilla.redhat.com/496135 missing dependency on PyXML Justin Forbes fixed both of these F-12 xen packaging bugs. https://bugzilla.redhat.com/531429 libvirt failed to disconnect and double free crashes on F12/rawhide with Xen Pasi Karkkainen requested a couple of libvirt xen driver fixes from 0.7.2 be backported to Fedora 12. From gianluca.cecchi at gmail.com Fri Oct 30 17:17:53 2009 From: gianluca.cecchi at gmail.com (Gianluca Cecchi) Date: Fri, 30 Oct 2009 18:17:53 +0100 Subject: [fedora-virt] duplicated console and details buttons In-Reply-To: <561c252c0910140044tb58325cv31a89d4c57fd550f@mail.gmail.com> References: <561c252c0910130902i19a76093h5b755b949458cf03@mail.gmail.com> <4AD4C1E8.50407@redhat.com> <561c252c0910140044tb58325cv31a89d4c57fd550f@mail.gmail.com> Message-ID: <561c252c0910301017j198488a0gcbe9db89e74c0fa1@mail.gmail.com> Hi, casually I have discovered what causes the anomaly, while installing a new f12 snapshot in a guest.... I'm now at my F11 host with: virt-manager-0.8.0-7.fc11.noarch libvirt-0.7.1-15.fc11.x86_64 qemu-0.11.0-9.fc11.x86_64 kernel-2.6.30.9-90.fc11.x86_64 When you start virt-manager and open a guest window, all is ok. Then if you select top-right "Full screen" icon, you get your guest in full screen mode. When you finish to work in full screen you select: view --> uncheck the full screen box You re-get the windowed mode of your guest, but with both the toolbar and the labels At this point if you close the guest window and reopen it, you always get the anomaly; instead if you close virt-manager and restart it, you re-get the initial correct behaviour. Not so harm, but just to follow up on my previous post. Bye, Gianluca On Wed, Oct 14, 2009 at 8:44 AM, Gianluca Cecchi wrote: > On Tue, Oct 13, 2009 at 8:07 PM, Cole Robinson wrote: > >> >> Hmm, those tabs should be hidden (they are for me on rawhide). What >> virt-manager and fedora version is this? >> >> - Cole >> >> > I'm on F11 x86_64 updated+ fedora-virt-preview repo. > virt-manager-0.8.0-7.fc11.noarch > libvirt-0.7.1-11.fc11.x86_64 > qemu-0.11.0-6.fc11.x86_64 > kernel-2.6.30.8-64.fc11.x86_64 > > Gianluca > -------------- next part -------------- An HTML attachment was scrubbed... URL: From crobinso at redhat.com Fri Oct 30 17:55:37 2009 From: crobinso at redhat.com (Cole Robinson) Date: Fri, 30 Oct 2009 13:55:37 -0400 Subject: [fedora-virt] duplicated console and details buttons In-Reply-To: <561c252c0910301017j198488a0gcbe9db89e74c0fa1@mail.gmail.com> References: <561c252c0910130902i19a76093h5b755b949458cf03@mail.gmail.com> <4AD4C1E8.50407@redhat.com> <561c252c0910140044tb58325cv31a89d4c57fd550f@mail.gmail.com> <561c252c0910301017j198488a0gcbe9db89e74c0fa1@mail.gmail.com> Message-ID: <4AEB2899.6030800@redhat.com> Gianluca Cecchi wrote: > Hi, > casually I have discovered what causes the anomaly, while installing a > new f12 snapshot in a guest.... > I'm now at my F11 host with: > virt-manager-0.8.0-7.fc11.noarch > libvirt-0.7.1-15.fc11.x86_64 > qemu-0.11.0-9.fc11.x86_64 > kernel-2.6.30.9-90.fc11.x86_64 > > When you start virt-manager and open a guest window, all is ok. > Then if you select top-right "Full screen" icon, you get your guest in > full screen mode. > When you finish to work in full screen you select: > view --> uncheck the full screen box > > You re-get the windowed mode of your guest, but with both the toolbar > and the labels > At this point if you close the guest window and reopen it, you always > get the anomaly; instead if you close virt-manager > and restart it, you re-get the initial correct behaviour. > > Not so harm, but just to follow up on my previous post. > Actually, there is a bug tracking this now where the same issue was discovered: https://bugzilla.redhat.com/show_bug.cgi?id=529657 Forgot to mention it in this thread, sorry. The issue is now fixed upstream though and this will be built for f12 + virt-preview soon. - Cole From ieidus at redhat.com Fri Oct 30 19:22:05 2009 From: ieidus at redhat.com (Izik Eidus) Date: Fri, 30 Oct 2009 21:22:05 +0200 Subject: [fedora-virt] disabling ksm by default In-Reply-To: <1256658916.25282.75.camel@blaa> References: <4AE628CA.1010807@redhat.com> <1256638369.25282.60.camel@blaa> <4AE6CA4D.9010604@redhat.com> <20091027123749.GA4530@redhat.com> <20091027144209.GA13805@linuxtx.org> <4AE71264.7000102@redhat.com> <1256658393.2803.31.camel@fedora64.linuxtx.org> <1256658916.25282.75.camel@blaa> Message-ID: <4AEB3CDD.4040903@redhat.com> On 10/27/2009 05:55 PM, Mark McLoughlin wrote: > On Tue, 2009-10-27 at 10:46 -0500, Justin M. Forbes wrote: > >> On Tue, 2009-10-27 at 17:31 +0200, Izik Eidus wrote: >> >>> On 10/27/2009 04:42 PM, Justin M. Forbes wrote: >>> >>> >>>> While I do understand what you are saying, I don't think it is worth making >>>> a kernel change for at this point in the cycle. Because ksm itself has a >>>> separate initscript, people who wish to use ksm will likely turn it on. >>>> This sets the max_kernel_pages to a reasonable value. People who are not >>>> interested enough to turn on the ksm service are probably not the kind of >>>> people who will be checking to see how effective ksm is at all. >>>> >>>> >>> To me it sound that users have no idea about this script, and ksm merge >>> to him just the zero pages of windows 7... >>> >>> My feeling is that 99% of the ppl in the world that will use it, would >>> just see the zero page merged and think "that is it..." >>> The current behaivor in fedora 12 is misleading the user (at least it >>> seems to me that it misslead that specific user) >>> >>> Btw we can set this value from userspace if we want "echo 0> >>> /sys/kernel/mm/ksm/run" >>> >> Where would we put that in userspace? Another init script? I just >> updated to documentation on the feature page (and hopefully in the final >> release notes) to say that ksm can be enabled by running 'sudo chkconfig >> ksm on' in effort to draw more attention to these scripts. >> >> The other option is to make the ksm init script default to on, and just >> leave ksmtuned off at system start. Since ksm is on by the default >> kernel, all we would be doing here is changing max_kernel_pages to a >> machine specific value. The ksm init script is part of the qemu >> package, so it would only be changed on systems doing virt anyway. >> > For Fedora 13, it'll be off by default in the kernel and the recommended > way of switching it on is with 'chkconfig ksm on' > > For Fedora 12, it's on by default in the kernel, 'chkconfig ksm on' just > changes max pages and the only way of disabling it is by manually > writing zero to /sys/kernel/mm/ksm/run > > IMHO, if we are to change anything for Fedora 12, we should move closer > to Fedora 13 behaviour, not further away from it - i.e. make it off by > default in the kernel, but I'm fine with delaying that until post GA > So what the result of this thread? you are delaying it to post GA? > We'll probably end up with this behaviour in F12 updates at some point > anyway when 2.6.32 is pulled in > > Cheers, > Mark. > >