From danken at redhat.com Sun Nov 1 06:42:41 2009 From: danken at redhat.com (Dan Kenigsberg) Date: Sun, 1 Nov 2009 08:42:41 +0200 Subject: [fedora-virt] disabling ksm by default In-Reply-To: <20091030150420.GA22371@linuxtx.org> References: <4AE628CA.1010807@redhat.com> <1256638369.25282.60.camel@blaa> <20091030100028.GA12068@amd.home.annexia.org> <20091030150420.GA22371@linuxtx.org> Message-ID: <20091101064240.GA20687@redhat.com> On Fri, Oct 30, 2009 at 10:04:20AM -0500, Justin M. Forbes wrote: > 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. Having `chkconfig ksm on` by default is great. I just want to see that the kernel thread, which ksm service controls, is not running until the service has started. As you say, when there are no qemu processes, the kernel thread has very low cost. But it also has zero value. So why have it running? From rjones at redhat.com Tue Nov 3 20:17:30 2009 From: rjones at redhat.com (Richard W.M. Jones) Date: Tue, 3 Nov 2009 20:17:30 +0000 Subject: [fedora-virt] Mount guest filesystems in the host Message-ID: <20091103201730.GA18808@amd.home.annexia.org> Little bit quiet on this list at the moment, so here's some news ... We just built a package called 'libguestfs-mount' in Rawhide which lets you mount virtual machines' filesystems on the host, using FUSE. http://koji.fedoraproject.org/koji/buildinfo?buildID=139621 Use it like this (as non-root): mkdir /tmp/mnt guestmount -a disk.img -m /dev/VG/Root --ro /tmp/mnt where you can replace 'disk.img' with any guest's disk image, and /dev/VG/Root with the root device in that guest. Or to get virt-inspector to do the hard work of looking in the guest and arranging the disks for you: guestmount $(virt-inspector --ro-fish MyGuest) /tmp/mnt where MyGuest is the name of the guest (as known by libvirt). And then you can just look into /tmp/mnt and browse the guest's disk using whatever text or graphical tools you want. To unmount do: fusermount -u /tmp/mnt 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 andres at verot.com Wed Nov 4 09:44:36 2009 From: andres at verot.com (=?ISO-8859-1?Q?Andr=E9s_Garc=EDa?=) Date: Wed, 04 Nov 2009 10:44:36 +0100 Subject: [fedora-virt] Network bridges in F12 beta Message-ID: <4AF14D04.1020602@verot.com> Hi, I think I read somewhere that with Fedora 12 there was going to be a way to configure virtual network bridges automagically. I have looked for it in virt-manager and network manager but it doesn't seem to be there, so I guess my memory is fooling me, right? So, if I have to do it manually, do the instructions at: http://wiki.libvirt.org/page/Networking#Bridged_networking_.28aka_.22shared_physical_device.22.29 still apply? Thanks, Andr?s From crobinso at redhat.com Wed Nov 4 14:25:25 2009 From: crobinso at redhat.com (Cole Robinson) Date: Wed, 04 Nov 2009 09:25:25 -0500 Subject: [fedora-virt] Network bridges in F12 beta In-Reply-To: <4AF14D04.1020602@verot.com> References: <4AF14D04.1020602@verot.com> Message-ID: <4AF18ED5.8080509@redhat.com> On 11/04/2009 04:44 AM, Andr?s Garc?a wrote: > Hi, > > I think I read somewhere that with Fedora 12 there was going to be a way > to configure > virtual network bridges automagically. > > I have looked for it in virt-manager and network manager but it doesn't > seem to be there, > so I guess my memory is fooling me, right? > Libvirt can configure bridges in f12, however it requires manually crafting XML and has no GUI support yet, so it is probably easiest to use the instructions you post below. - Cole > So, if I have to do it manually, do the instructions at: > > http://wiki.libvirt.org/page/Networking#Bridged_networking_.28aka_.22shared_physical_device.22.29 > > still apply? > > Thanks, > Andr?s > > _______________________________________________ > Fedora-virt mailing list > Fedora-virt at redhat.com > https://www.redhat.com/mailman/listinfo/fedora-virt From sbonnevi at redhat.com Wed Nov 4 17:20:57 2009 From: sbonnevi at redhat.com (Steve Bonneville) Date: Wed, 4 Nov 2009 11:20:57 -0600 Subject: [fedora-virt] Network bridges in F12 beta In-Reply-To: <20091104170012.274F98E03E5@hormel.redhat.com> References: <20091104170012.274F98E03E5@hormel.redhat.com> Message-ID: <20091104172057.GD31915@sbonnevi-lt.msp.redhat.com> Cole Robinson wrote: > On 11/04/2009 04:44 AM, Andr?s Garc?a wrote: > > Hi, > > > > I think I read somewhere that with Fedora 12 there was going to be a way > > to configure > > virtual network bridges automagically. > > > > I have looked for it in virt-manager and network manager but it doesn't > > seem to be there, > > so I guess my memory is fooling me, right? > > > > Libvirt can configure bridges in f12, however it requires manually > crafting XML and has no GUI support yet, so it is probably easiest to > use the instructions you post below. Do you have a reference to how the bridges would be crafted in libvirt XML? Which version of libvirt is required? -- Steve -- Steven Bonneville Curriculum Manager (Linux Team Lead) Red Hat | Global Learning Services Phone: +1-612-638-0507 gpg: 1024D/221D06FF 68B1 3E66 A351 6485 B9AF 24D8 3DF5 B50B 221D 06FF From crobinso at redhat.com Wed Nov 4 17:28:16 2009 From: crobinso at redhat.com (Cole Robinson) Date: Wed, 04 Nov 2009 12:28:16 -0500 Subject: [fedora-virt] Network bridges in F12 beta In-Reply-To: <20091104172057.GD31915@sbonnevi-lt.msp.redhat.com> References: <20091104170012.274F98E03E5@hormel.redhat.com> <20091104172057.GD31915@sbonnevi-lt.msp.redhat.com> Message-ID: <4AF1B9B0.4090409@redhat.com> On 11/04/2009 12:20 PM, Steve Bonneville wrote: > Cole Robinson wrote: >> On 11/04/2009 04:44 AM, Andr?s Garc?a wrote: >>> Hi, >>> >>> I think I read somewhere that with Fedora 12 there was going to be a way >>> to configure >>> virtual network bridges automagically. >>> >>> I have looked for it in virt-manager and network manager but it doesn't >>> seem to be there, >>> so I guess my memory is fooling me, right? >>> >> >> Libvirt can configure bridges in f12, however it requires manually >> crafting XML and has no GUI support yet, so it is probably easiest to >> use the instructions you post below. > > Do you have a reference to how the bridges would be crafted in libvirt > XML? Which version of libvirt is required? > You need libvirt 0.7.0 or later. As for docs, virtually nonexistent at the moment, your best bet is to check out the netcf source and take a look at the tests/ directory: https://fedorahosted.org/netcf/ - Cole From andres at verot.com Wed Nov 4 18:36:53 2009 From: andres at verot.com (=?ISO-8859-1?Q?Andr=E9s_Garc=EDa?=) Date: Wed, 04 Nov 2009 19:36:53 +0100 Subject: [fedora-virt] Host lock up in F11 Message-ID: <4AF1C9C5.3000609@verot.com> Hi, I sometimes have my F11 virtual machine server crash with leaving me with nothing but blinking keyboard lights. Searching for a solution I found this: http://www.savelono.com/linux/update-using-multiple-interfaces-with-kvm-and-xen.html Could someone please tell me, if the post is right and whether the problem has been solved in F12. Thanks, Andres From andres at verot.com Thu Nov 5 15:48:29 2009 From: andres at verot.com (=?ISO-8859-1?Q?Andr=E9s_Garc=EDa?=) Date: Thu, 05 Nov 2009 16:48:29 +0100 Subject: [fedora-virt] Connect from remote host Message-ID: <4AF2F3CD.6040908@verot.com> Hi, I configured a Fedora 12 beta system to accept connections from other system as explained here: http://virt-manager.et.redhat.com/page/RemoteDigest To test it I connected using virt-manager in that system to itself and it worked, but when I tried from a F11 virt-manager in another system it fails, I also tried connecting with virsh: # virsh virsh # connect qemu+tcp://192.168.0.43 It tells me there is no route to host despite being able to ping 192.168.0.43. Any ideas what could be wrong? F12 has a firewall I am not aware of? Thanks, Andres Garcia From crobinso at redhat.com Thu Nov 5 16:41:05 2009 From: crobinso at redhat.com (Cole Robinson) Date: Thu, 05 Nov 2009 11:41:05 -0500 Subject: [fedora-virt] Connect from remote host In-Reply-To: <4AF2F3CD.6040908@verot.com> References: <4AF2F3CD.6040908@verot.com> Message-ID: <4AF30021.1030809@redhat.com> On 11/05/2009 10:48 AM, Andr?s Garc?a wrote: > Hi, > > I configured a Fedora 12 beta system to accept connections from other > system as explained > here: > > http://virt-manager.et.redhat.com/page/RemoteDigest > > To test it I connected using virt-manager in that system to itself and > it worked, but when I tried > from a F11 virt-manager in another system it fails, I also tried > connecting with virsh: > > # virsh > virsh # connect qemu+tcp://192.168.0.43 Does changing this to qemu+tcp://192.168.0.43/system make any difference? > > It tells me there is no route to host despite being able to ping > 192.168.0.43. > An actual copy of the error output could help. Also stick LIBVIRT_DEBUG=1 before the virsh command. > Any ideas what could be wrong? F12 has a firewall I am not aware of? > Turn it off with service iptables stop and see if it helps at all. > Thanks, > Andres Garcia > - Cole From tom.horsley at att.net Thu Nov 5 17:03:01 2009 From: tom.horsley at att.net (Tom Horsley) Date: Thu, 5 Nov 2009 12:03:01 -0500 Subject: [fedora-virt] ubuntu 9.10 KVM on fedora 11 question Message-ID: <20091105120301.4f5eac1f@tomh> I have some very weird behavior with my new ubuntu 9.10 virtual machines running as KVMs under fedora 11. I run testbeds on lots of different linuxii in ssh sessions to virtual machines, and the ssh session just stops randomly when talking to ubuntu 9.10 (no problem with 9.04, or 8.10, no problem with various fedora versions or openSUSE versions). I can still ssh into the KVM with a new ssh connection, but the original connection is frozen up, it never even times out an disconnects, it acts like it is in the middle of talking and is just waiting for the next thing to happen. Does this strike a familiar note to anyone? I've tried both virtio and emulated NIC modes, and see the same results. From andres at verot.com Thu Nov 5 17:04:51 2009 From: andres at verot.com (=?ISO-8859-1?Q?Andr=E9s_Garc=EDa?=) Date: Thu, 05 Nov 2009 18:04:51 +0100 Subject: [fedora-virt] Connect from remote host In-Reply-To: <4AF30021.1030809@redhat.com> References: <4AF2F3CD.6040908@verot.com> <4AF30021.1030809@redhat.com> Message-ID: <4AF305B3.1030400@verot.com> Hi, > Turn it off with service iptables stop and see if it helps at all. > > That did it, sorry if this is a silly question but iptables have always been a bit of a mistery to me, may I disable them without any ill effects? The office already has a firewall, so I don't really need it for security reasons. Thanks, Andr?s Garc?a From crobinso at redhat.com Thu Nov 5 18:43:14 2009 From: crobinso at redhat.com (Cole Robinson) Date: Thu, 05 Nov 2009 13:43:14 -0500 Subject: [fedora-virt] Connect from remote host In-Reply-To: <4AF305B3.1030400@verot.com> References: <4AF2F3CD.6040908@verot.com> <4AF30021.1030809@redhat.com> <4AF305B3.1030400@verot.com> Message-ID: <4AF31CC2.2050803@redhat.com> On 11/05/2009 12:04 PM, Andr?s Garc?a wrote: > Hi, >> Turn it off with service iptables stop and see if it helps at all. >> >> > > That did it, sorry if this is a silly question but iptables have always > been a bit of a mistery to > me, may I disable them without any ill effects? The office already has a > firewall, so I don't really > need it for security reasons. > There are potential security implications, but if you are using a desktop machine (as in, this isn't a production server) already behind a corporate firewall, you probably have nothing to worry about. - Cole From psj at familyjenner.co.uk Thu Nov 5 20:40:40 2009 From: psj at familyjenner.co.uk (Paul Jenner) Date: Thu, 05 Nov 2009 20:40:40 +0000 Subject: [fedora-virt] Re: [libvirt] F12beta guest will not connect to LAN via bridge In-Reply-To: <4AF24BCE.7040405@verizon.net> References: <4AEFC0F4.9070405@verizon.net> <4AF1ABA8.8090004@verizon.net> <4AF24BCE.7040405@verizon.net> Message-ID: <1257453640.2302.1.camel@82-71-88-77.dsl.in-addr.zen.co.uk> Hi Gerry. On Wed, 2009-11-04 at 22:51 -0500, Gerry Reno wrote: > Gerry Reno wrote: > > Gerry Reno wrote: > >> Setup: > >> HOST: F11 x86_64 > >> GUEST: F12beta x86_64 > >> > >> I have a number of guests setup on this host and all connect without > >> a problem with the 'br0' bridge on the host. I created a new guest > >> and loaded F12beta on it. I defined it to also use 'br0' networking > >> but this guest cannot ping anything on the LAN. All the other guests > >> have no problems communicating to addresses on the LAN. > > The only thing I can find is that the new F12beta guest does not have > > 'bridge' module loaded in kernel as compared with a working existing > > F11 guest. I tried loading the 'bridge' module but this did not make > > any difference after restarting the network. I then tried rebooting > > the guest but then the 'bridge' module is not loaded again. Is the > > 'bridge' module necessary? I thought I had read where this was no > > longer necessary for guest bridging. > > > The only way I can get bridged networking to work in the F12beta guest > is by selecting "Hypervisor default" when adding Network device. If I > select "virtio" it fails. This sounds like it should be raised as a Fedora bug? Paul -- Paul Jenner From jmforbes at linuxtx.org Thu Nov 5 22:44:49 2009 From: jmforbes at linuxtx.org (Justin M. Forbes) Date: Thu, 5 Nov 2009 16:44:49 -0600 Subject: [fedora-virt] Re: [libvirt] F12beta guest will not connect to LAN via bridge In-Reply-To: <1257453640.2302.1.camel@82-71-88-77.dsl.in-addr.zen.co.uk> References: <4AEFC0F4.9070405@verizon.net> <4AF1ABA8.8090004@verizon.net> <4AF24BCE.7040405@verizon.net> <1257453640.2302.1.camel@82-71-88-77.dsl.in-addr.zen.co.uk> Message-ID: <20091105224449.GA5137@linuxtx.org> On Thu, Nov 05, 2009 at 08:40:40PM +0000, Paul Jenner wrote: > > This sounds like it should be raised as a Fedora bug? Indeed it has been. Bug 532710 Justin From andres at verot.com Fri Nov 6 15:44:02 2009 From: andres at verot.com (=?ISO-8859-1?Q?Andr=E9s_Garc=EDa?=) Date: Fri, 06 Nov 2009 16:44:02 +0100 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: <4AF44442.8010903@verot.com> Hi, > ksm is disabled by default currently > > Does that mean its use is deprecated? Andres From gmaddock at futuremetals.com Fri Nov 6 15:50:43 2009 From: gmaddock at futuremetals.com (Gerry Maddock) Date: Fri, 6 Nov 2009 10:50:43 -0500 Subject: [fedora-virt] polkit-gnome-authorization question Message-ID: An HTML attachment was scrubbed... URL: From jmforbes at linuxtx.org Fri Nov 6 16:19:11 2009 From: jmforbes at linuxtx.org (Justin M. Forbes) Date: Fri, 6 Nov 2009 10:19:11 -0600 Subject: [fedora-virt] disabling ksm by default In-Reply-To: <4AF44442.8010903@verot.com> References: <4AE628CA.1010807@redhat.com> <1256638369.25282.60.camel@blaa> <4AF44442.8010903@verot.com> Message-ID: <20091106161911.GB5137@linuxtx.org> On Fri, Nov 06, 2009 at 04:44:02PM +0100, Andr?s Garc?a wrote: >> ksm is disabled by default currently >> >> > > Does that mean its use is deprecated? > No, it is not and ksm is enabled by default on F12 now. Thanks, Justin From gmaddock at futuremetals.com Fri Nov 6 18:32:02 2009 From: gmaddock at futuremetals.com (Gerry Maddock) Date: Fri, 6 Nov 2009 13:32:02 -0500 Subject: [fedora-virt] virtual machine manager - rename virtual machine In-Reply-To: References: Message-ID: An HTML attachment was scrubbed... URL: From tom.horsley at att.net Fri Nov 6 18:38:08 2009 From: tom.horsley at att.net (Tom Horsley) Date: Fri, 6 Nov 2009 13:38:08 -0500 Subject: [fedora-virt] virtual machine manager - rename virtual machine In-Reply-To: References: Message-ID: <20091106133808.6651e8de@tomh> On Fri, 6 Nov 2009 13:32:02 -0500 Gerry Maddock wrote: > Is there a way I can rename an existing virtual machine as seen from virtual machine manager? I don't see a way to do it from the Virtual Machine manager, but I'd rather edit a file anyway. One way would be something like: virsh dumpxml oldname > newname.xml virsh undefine oldname edit newname.xml to change name embedded in the xml virsh define newname.xml of course, that is probably not a good idea while the machine is running. From gmaddock at futuremetals.com Fri Nov 6 19:05:28 2009 From: gmaddock at futuremetals.com (Gerry Maddock) Date: Fri, 6 Nov 2009 14:05:28 -0500 Subject: [fedora-virt] virtual machine manager - rename virtual machine In-Reply-To: <20091106133808.6651e8de@tomh> References: , <20091106133808.6651e8de@tomh> Message-ID: An HTML attachment was scrubbed... URL: From markmc at redhat.com Mon Nov 9 14:00:40 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Mon, 09 Nov 2009 14:00:40 +0000 Subject: [fedora-virt] Network bridges in F12 beta In-Reply-To: <4AF18ED5.8080509@redhat.com> References: <4AF14D04.1020602@verot.com> <4AF18ED5.8080509@redhat.com> Message-ID: <1257775240.3007.28.camel@blaa> On Wed, 2009-11-04 at 09:25 -0500, Cole Robinson wrote: > On 11/04/2009 04:44 AM, Andr?s Garc?a wrote: > > Hi, > > > > I think I read somewhere that with Fedora 12 there was going to be a way > > to configure > > virtual network bridges automagically. > > > > I have looked for it in virt-manager and network manager but it doesn't > > seem to be there, > > so I guess my memory is fooling me, right? > > > > Libvirt can configure bridges in f12, however it requires manually > crafting XML and has no GUI support yet, so it is probably easiest to > use the instructions you post below. > > - Cole > > > So, if I have to do it manually, do the instructions at: > > > > http://wiki.libvirt.org/page/Networking#Bridged_networking_.28aka_.22shared_physical_device.22.29 > > > > still apply? It'd be really nice to have a version of those instructions that uses e.g. virsh iface-define Cheers, Mark. From markmc at redhat.com Mon Nov 9 14:12:47 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Mon, 09 Nov 2009 14:12:47 +0000 Subject: [fedora-virt] Host lock up in F11 In-Reply-To: <4AF1C9C5.3000609@verot.com> References: <4AF1C9C5.3000609@verot.com> Message-ID: <1257775967.3007.31.camel@blaa> On Wed, 2009-11-04 at 19:36 +0100, Andr?s Garc?a wrote: > Hi, > > I sometimes have my F11 virtual machine server crash with leaving me > with nothing but > blinking keyboard lights. > > Searching for a solution I found this: > > http://www.savelono.com/linux/update-using-multiple-interfaces-with-kvm-and-xen.html > > Could someone please tell me, if the post is right and whether the > problem has been solved > in F12. This sounds like: https://bugzilla.redhat.com/483532 I can't find any reference to it being fixed, but it may just have gone away with newer kernels Best bet is probably to hook up a serial console so you can get the full oops info Cheers, Mark. From markmc at redhat.com Mon Nov 9 14:15:17 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Mon, 09 Nov 2009 14:15:17 +0000 Subject: [fedora-virt] ubuntu 9.10 KVM on fedora 11 question In-Reply-To: <20091105120301.4f5eac1f@tomh> References: <20091105120301.4f5eac1f@tomh> Message-ID: <1257776117.3007.33.camel@blaa> On Thu, 2009-11-05 at 12:03 -0500, Tom Horsley wrote: > I have some very weird behavior with my new ubuntu 9.10 > virtual machines running as KVMs under fedora 11. > > I run testbeds on lots of different linuxii in ssh sessions > to virtual machines, and the ssh session just stops randomly > when talking to ubuntu 9.10 (no problem with 9.04, or 8.10, > no problem with various fedora versions or openSUSE versions). > > I can still ssh into the KVM with a new ssh connection, but > the original connection is frozen up, it never even times > out an disconnects, it acts like it is in the middle of > talking and is just waiting for the next thing to happen. > > Does this strike a familiar note to anyone? I've tried > both virtio and emulated NIC modes, and see the same results. I don't recall hearing about this bug before I'd be curious as to whether you can reproduce in with the Fedora 12 bits from virt-preview. A tcpdump might help too Also might be worth asking Ubuntu developers, it could just be a guest bug Cheers, Mark. From tom.horsley at att.net Mon Nov 9 17:05:50 2009 From: tom.horsley at att.net (Tom Horsley) Date: Mon, 9 Nov 2009 12:05:50 -0500 Subject: [fedora-virt] ubuntu 9.10 KVM on fedora 11 question In-Reply-To: <1257776117.3007.33.camel@blaa> References: <20091105120301.4f5eac1f@tomh> <1257776117.3007.33.camel@blaa> Message-ID: <20091109120550.5857c4f2@tomh> On Mon, 09 Nov 2009 14:15:17 +0000 Mark McLoughlin wrote: > On Thu, 2009-11-05 at 12:03 -0500, Tom Horsley wrote: > > ... > > I can still ssh into the KVM with a new ssh connection, but > > the original connection is frozen up, it never even times > > out an disconnects, it acts like it is in the middle of > > talking and is just waiting for the next thing to happen. > > > > Does this strike a familiar note to anyone? I've tried > > both virtio and emulated NIC modes, and see the same results. > > I don't recall hearing about this bug before Well, ubuntu 9.10 was just released, and it is the only system I've had this problem with. > I'd be curious as to whether you can reproduce in with the Fedora 12 > bits from virt-preview. A tcpdump might help too Yea, I'm curious about those too, but I'm not really in the position to switch the host software around a lot (though I may indeed go to fedora 12 in the not too distant future). Not sure if I have enough disk space to record a tcpdump :-). > Also might be worth asking Ubuntu developers, it could just be a guest > bug Yea, I've asked over on ubuntuforums, but no response there as yet. One thing I did was install ubuntu 9.10 on some real hardware, and that system has run the testbeds several times with no problems, so it does seem to be something unique to the KVM installs. From tom.horsley at att.net Mon Nov 9 20:17:10 2009 From: tom.horsley at att.net (Tom Horsley) Date: Mon, 9 Nov 2009 15:17:10 -0500 Subject: [fedora-virt] ubuntu 9.10 KVM on fedora 11 question In-Reply-To: <20091109120550.5857c4f2@tomh> References: <20091105120301.4f5eac1f@tomh> <1257776117.3007.33.camel@blaa> <20091109120550.5857c4f2@tomh> Message-ID: <20091109151710.6a9cd242@tomh> On Mon, 9 Nov 2009 12:05:50 -0500 Tom Horsley wrote: > On Mon, 09 Nov 2009 14:15:17 +0000 > Mark McLoughlin wrote: > > > On Thu, 2009-11-05 at 12:03 -0500, Tom Horsley wrote: > > > ... > > > I can still ssh into the KVM with a new ssh connection, but > > > the original connection is frozen up, it never even times > > > out an disconnects, it acts like it is in the middle of > > > talking and is just waiting for the next thing to happen. Woa! Just happened to notice new symptom. I had an ssh session to one of these ubuntu KVM machines open at the same time I had watched it boot in the console window I opened from virt-manager. Not only did the ssh session freeze, but the console window froze at the same time. I could still ssh in with a new ssh session, but I'm not sure what would take out an ssh session and the console at the same time while leaving the KVM machine up. From gene at czarc.net Mon Nov 9 20:36:16 2009 From: gene at czarc.net (Gene Czarcinski) Date: Mon, 9 Nov 2009 15:36:16 -0500 Subject: [fedora-virt] Network bridges in F12 beta In-Reply-To: <1257775240.3007.28.camel@blaa> References: <4AF14D04.1020602@verot.com> <4AF18ED5.8080509@redhat.com> <1257775240.3007.28.camel@blaa> Message-ID: <200911091536.16444.gene@czarc.net> On Monday 09 November 2009 09:00:40 Mark McLoughlin wrote: > On Wed, 2009-11-04 at 09:25 -0500, Cole Robinson wrote: > > On 11/04/2009 04:44 AM, Andr?s Garc?a wrote: > > > Hi, > > > > > > I think I read somewhere that with Fedora 12 there was going to be a > > > way to configure > > > virtual network bridges automagically. > > > > > > I have looked for it in virt-manager and network manager but it doesn't > > > seem to be there, > > > so I guess my memory is fooling me, right? > > > > Libvirt can configure bridges in f12, however it requires manually > > crafting XML and has no GUI support yet, so it is probably easiest to > > use the instructions you post below. > > > > - Cole > > > > > So, if I have to do it manually, do the instructions at: > > > > > > http://wiki.libvirt.org/page/Networking#Bridged_networking_.28aka_.22sh > > >ared_physical_device.22.29 > > > > > > still apply? > > It'd be really nice to have a version of those instructions that uses > e.g. virsh iface-define It sure would be nice if there was more documentation available about all of the "stuff" that can be done with virsh ... the man page leaves a lot unsaid! The implication that you could use "virsh iface-define" to setup a bridge definition on a system sounds very powerful. However, I thought that virsh was something you used with virtual stuff ... not real hardware. Setting up a bridge involves definitions for the host system ... not a virtual guest. Creating xml with a text editor is not something I would look forward to just to do the bridge definition. Now if you want to prod the NetworkManager folks to add bridge support sooner rather than later, that is something else. Gene From eb30750 at gmail.com Tue Nov 10 01:03:00 2009 From: eb30750 at gmail.com (Paul Lambert) Date: Mon, 9 Nov 2009 20:03:00 -0500 Subject: [fedora-virt] Sound status on VM? Message-ID: Does sound now work with a VM? I am using rawhide x86_64 for the host and FE-11 x86 for the guest. I also have a Windows XP VM. The VM hardware shows es1370. The host sound plays mp3s using either Totem or Rhythmbox just fine but no sound from either VM. Thanks Paul -------------- next part -------------- An HTML attachment was scrubbed... URL: From markmc at redhat.com Tue Nov 10 08:03:47 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Tue, 10 Nov 2009 08:03:47 +0000 Subject: [fedora-virt] ubuntu 9.10 KVM on fedora 11 question In-Reply-To: <20091109151710.6a9cd242@tomh> References: <20091105120301.4f5eac1f@tomh> <1257776117.3007.33.camel@blaa> <20091109120550.5857c4f2@tomh> <20091109151710.6a9cd242@tomh> Message-ID: <1257840227.2888.19.camel@blaa> On Mon, 2009-11-09 at 15:17 -0500, Tom Horsley wrote: > On Mon, 9 Nov 2009 12:05:50 -0500 > Tom Horsley wrote: > > > On Mon, 09 Nov 2009 14:15:17 +0000 > > Mark McLoughlin wrote: > > > > > On Thu, 2009-11-05 at 12:03 -0500, Tom Horsley wrote: > > > > ... > > > > I can still ssh into the KVM with a new ssh connection, but > > > > the original connection is frozen up, it never even times > > > > out an disconnects, it acts like it is in the middle of > > > > talking and is just waiting for the next thing to happen. > > Woa! Just happened to notice new symptom. I had an ssh session > to one of these ubuntu KVM machines open at the same time I had > watched it boot in the console window I opened from virt-manager. > Not only did the ssh session freeze, but the console window froze > at the same time. I could still ssh in with a new ssh session, but > I'm not sure what would take out an ssh session and the console > at the same time while leaving the KVM machine up. Are you running virt-manager on the same machine as the guests? i.e. is virt-manager connected to qemu:///system ? If it's all local, the VNC connection is simply a localhost TCP socket between virt-manager and qemu, whereas the ssh is a TCP socket over virbr0/vnet0 ethernet to the guest sshd. I can't imagine how both would lock up at the same time while qemu is still happily processing packets to form a new ssh connection. Cheers, Mark. From markmc at redhat.com Tue Nov 10 08:05:45 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Tue, 10 Nov 2009 08:05:45 +0000 Subject: [fedora-virt] Sound status on VM? In-Reply-To: References: Message-ID: <1257840345.2888.21.camel@blaa> On Mon, 2009-11-09 at 20:03 -0500, Paul Lambert wrote: > Does sound now work with a VM? I am using rawhide x86_64 for the host > and FE-11 x86 for the guest. I also have a Windows XP VM. The VM > hardware shows es1370. The host sound plays mp3s using either Totem > or Rhythmbox just fine but no sound from either VM. We basically don't have a usable sound backend at the moment. The plan is to tunnel sound over VNC: https://bugzilla.redhat.com/508317 Cheers, Mark. From tom.horsley at att.net Tue Nov 10 12:40:34 2009 From: tom.horsley at att.net (Tom Horsley) Date: Tue, 10 Nov 2009 07:40:34 -0500 Subject: [fedora-virt] ubuntu 9.10 KVM on fedora 11 question In-Reply-To: <1257840227.2888.19.camel@blaa> References: <20091105120301.4f5eac1f@tomh> <1257776117.3007.33.camel@blaa> <20091109120550.5857c4f2@tomh> <20091109151710.6a9cd242@tomh> <1257840227.2888.19.camel@blaa> Message-ID: <20091110074034.7685db4e@tomh> On Tue, 10 Nov 2009 08:03:47 +0000 Mark McLoughlin wrote: > > Woa! Just happened to notice new symptom. I had an ssh session > > to one of these ubuntu KVM machines open at the same time I had > > watched it boot in the console window I opened from virt-manager. > > Not only did the ssh session freeze, but the console window froze > > at the same time. I could still ssh in with a new ssh session, but > > I'm not sure what would take out an ssh session and the console > > at the same time while leaving the KVM machine up. > > Are you running virt-manager on the same machine as the guests? i.e. is > virt-manager connected to qemu:///system ? > > If it's all local, the VNC connection is simply a localhost TCP socket > between virt-manager and qemu, whereas the ssh is a TCP socket over > virbr0/vnet0 ethernet to the guest sshd. I can't imagine how both would > lock up at the same time while qemu is still happily processing packets > to form a new ssh connection. Virt-manager was running local on the host machine, but displaying remotely to my desktop over a forwarded X connection. The ssh session was from my desktop directly to the virtual machine (all the VMs are setup using a bridge network). It is indeed weird. Other versions of linux running in VMs configured pretty much the same way on the same host don't have these problems. The same ubuntu 9.10 on real hardware doesn't have the problem. Worse yet, I installed some ubuntu 9.10 virtual machines at home on a very similar (but much less loaded) KVM setup as I have at work, and they have never shown the problem. I guess there are just some things man was not meant to know :-). I'll see what happens when we install fedora 12 on the host after the official release. From pasik at iki.fi Tue Nov 10 17:12:55 2009 From: pasik at iki.fi (Pasi =?iso-8859-1?Q?K=E4rkk=E4inen?=) Date: Tue, 10 Nov 2009 19:12:55 +0200 Subject: [fedora-virt] [Xen-devel] [ANNOUNCE] Xen 3.4.2 released Message-ID: <20091110171255.GZ16033@reaktio.net> ----- Forwarded message from Keir Fraser ----- From: Keir Fraser To: "xen-devel at lists.xensource.com" , xen-users at lists.xensource.com Cc: Date: Tue, 10 Nov 2009 16:38:59 +0000 Subject: [Xen-devel] [ANNOUNCE] Xen 3.4.2 released Folks, Xen 3.4.2 is the latest maintenance release in the 3.4 stable branch. There are a range of bug fixes since 3.4.1, and we recommend users to upgrade. The source repository can be downloaded from: http://xenbits.xensource.com/xen-3.4-testing.hg The release is tagged 'RELEASE-3.4.2'. Alternatively source tarballs can be downloaded from: http://www.xen.org/products/xen_source.html -- Keir _______________________________________________ Xen-devel mailing list Xen-devel at lists.xensource.com http://lists.xensource.com/xen-devel ----- End forwarded message ----- From rjones at redhat.com Tue Nov 10 17:31:10 2009 From: rjones at redhat.com (Richard W.M. Jones) Date: Tue, 10 Nov 2009 17:31:10 +0000 Subject: [fedora-virt] ubuntu 9.10 KVM on fedora 11 question In-Reply-To: <20091105120301.4f5eac1f@tomh> References: <20091105120301.4f5eac1f@tomh> Message-ID: <20091110173110.GA24191@amd.home.annexia.org> On Thu, Nov 05, 2009 at 12:03:01PM -0500, Tom Horsley wrote: > I have some very weird behavior with my new ubuntu 9.10 > virtual machines running as KVMs under fedora 11. > > I run testbeds on lots of different linuxii in ssh sessions > to virtual machines, and the ssh session just stops randomly > when talking to ubuntu 9.10 (no problem with 9.04, or 8.10, > no problem with various fedora versions or openSUSE versions). Do the guests have libvirtd running inside them? libvirtd tends to create a default network (192.168.122.*) which conflicts with the host virbr0 network. This usually breaks networking completely -- I've not seen is cause hangs on an existing working network, unless ssh is hanging while something is starting up libvirtd. https://bugzilla.redhat.com/show_bug.cgi?id=509458 If not that, then these problems are often caused by specific classes of packets getting through (eg. packets that are too large, have broken checksums etc). So as Mark said, a tcpdump is vital to solve the issue. 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 rjones at redhat.com Tue Nov 10 17:35:19 2009 From: rjones at redhat.com (Richard W.M. Jones) Date: Tue, 10 Nov 2009 17:35:19 +0000 Subject: [fedora-virt] ubuntu 9.10 KVM on fedora 11 question In-Reply-To: <20091110173110.GA24191@amd.home.annexia.org> References: <20091105120301.4f5eac1f@tomh> <20091110173110.GA24191@amd.home.annexia.org> Message-ID: <20091110173519.GB24191@amd.home.annexia.org> On Tue, Nov 10, 2009 at 05:31:10PM +0000, Richard W.M. Jones wrote: > of packets getting through (eg. packets that are too large, have ... not getting through ... 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 tom.horsley at att.net Tue Nov 10 18:25:28 2009 From: tom.horsley at att.net (Tom Horsley) Date: Tue, 10 Nov 2009 13:25:28 -0500 Subject: [fedora-virt] ubuntu 9.10 KVM on fedora 11 question In-Reply-To: <20091110173110.GA24191@amd.home.annexia.org> References: <20091105120301.4f5eac1f@tomh> <20091110173110.GA24191@amd.home.annexia.org> Message-ID: <20091110132528.6f76c57c@tomh> On Tue, 10 Nov 2009 17:31:10 +0000 Richard W.M. Jones wrote: > So as Mark said, a tcpdump is vital to solve > the issue. I still have no ideas why the virt-viewer console stopped working at the same time, but going through logs on the guest machines showed $#@! NetworkManager deciding it needed to do a brand new DHCPDISCOVER for the network connection which was already up, resulting in the IP address changing out from under my original connection. After much cursing and poking of sticks at NetworkManager I think the connection is now staying up. Why my "real hardware" ubuntu 9.10 machine doesn't suffer from the same problem, I don't know. From rjones at redhat.com Tue Nov 10 22:31:59 2009 From: rjones at redhat.com (Richard W.M. Jones) Date: Tue, 10 Nov 2009 22:31:59 +0000 Subject: [fedora-virt] ANNOUNCE: libguestfs 1.0.78 released Message-ID: <20091110223159.GA25441@amd.home.annexia.org> I'm pleased to announce the release of libguestfs 1.0.78. Libguestfs is a library for accessing and modifying virtual machine disk images. Home page: http://libguestfs.org/ Source: http://libguestfs.org/download/ Binaries: http://libguestfs.org/FAQ.html#binaries A Fedora build is available in Koji here: http://koji.fedoraproject.org/koji/packageinfo?packageID=8391 (These release notes cover all the significant changes since the last announcement which was for 1.0.67, 2 months ago). New features: - FUSE support so you can mount guest filesystems in the host: http://rwmj.wordpress.com/2009/11/03/browsing-guests-using-fuse/ - Support for btrfs, gfs, gfs2, hfs, hfs+, nilfs2, jfs, reiserfs, xfs: http://rwmj.wordpress.com/2009/11/08/filesystem-metadata-overhead/ - Support for huge (multi-exabyte) sparse virtual disks: http://rwmj.wordpress.com/2009/11/04/petabytes-exabytes-why-not/ http://git.et.redhat.com/?p=libguestfs.git;a=commit;h=5ce72e039ca332ba19bb9122b7c93d257e745bb5 - New partitioning API, supports GPT and more: http://libguestfs.org/guestfs.3.html#guestfs_part_add http://git.et.redhat.com/?p=libguestfs.git;a=commit;h=b1e1ca2f74a921b3f784537d59c617df29ea1d60 - New tool: virt-ls http://rwmj.wordpress.com/2009/10/19/new-tool-virt-ls/ - New tool: virt-tar http://rwmj.wordpress.com/2009/10/19/new-tool-virt-tar/ - New tool: virt-edit http://rwmj.wordpress.com/2009/09/23/virt-edit/ - New tool: virt-rescue http://rwmj.wordpress.com/2009/09/22/virt-rescue/ - Windows Registry support, tools and library: http://rwmj.wordpress.com/2009/10/29/virt-win-reg-get-at-the-windows-registry-in-your-windows-guests/ http://rwmj.wordpress.com/2009/10/28/libhivex-windows-registry-hive-extractor-library/ - OCaml bindings for virt-inspector - RELAX NG schema for virt-inspector - New APIs: utimens, vfs_type, truncate, truncate_size, lchown, lstatlist, lxattrlist, readlinklist, case_sensitive_path, find0, mkfs_b, mke2journal, and more ... - New program: OCaml viewer http://rwmj.wordpress.com/2009/09/29/graphical-virt-df/ - Allow stdout to be redirected when running guestfish remotely (Matt Booth). - Remove requirement for vmchannel support in qemu (horray!) and the tricky main loop code. Bug fixes: (Too many to list here, see: http://git.et.redhat.com/?p=libguestfs.git;a=log) Thanks to: Jim Meyering, Matt Booth and Charles Duffy for lots of bugfixes, patches and testing. 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 danken at redhat.com Wed Nov 11 21:37:37 2009 From: danken at redhat.com (Dan Kenigsberg) Date: Wed, 11 Nov 2009 23:37:37 +0200 Subject: [fedora-virt] disabling ksm by default In-Reply-To: <4AEB3CDD.4040903@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> <4AEB3CDD.4040903@redhat.com> Message-ID: <20091111213736.GA1445@csdan.redhat.com> On Fri, Oct 30, 2009 at 09:22:05PM +0200, Izik Eidus wrote: > 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? kernel is kept with ksm_run = KSM_RUN_MERGE by deafult (different from upstream, God forbid!), but init.d/ksm runs by default and sets max_kernel_pages to a reasonable value. http://gitorious.org/ksm-control-scripts/ksm-control-scripts/commit/0a242288d0b7f3606e68ce2096e467e337d22c0a From dlbewley at lib.ucdavis.edu Sun Nov 15 21:10:22 2009 From: dlbewley at lib.ucdavis.edu (Dale Bewley) Date: Sun, 15 Nov 2009 13:10:22 -0800 (PST) Subject: [fedora-virt] bridged guest networking in f12 with netcf In-Reply-To: <1520168297.823191258319345201.JavaMail.root@zebra.lib.ucdavis.edu> Message-ID: <1205932802.823231258319422164.JavaMail.root@zebra.lib.ucdavis.edu> Hello all, I started a blog post about experimenting with netcf and configuring a guest to share the same subnet as the host. It turned into something a bit longer and more rambling of course. * http://tofu.org/drupal/node/86 At any rate I ran into a few issues while using: virt-manager-0.8.0-7.fc12.noarch netcf-0.1.4-1.fc12.x86_64 libvirt-0.7.1-15.fc12.x86_64 = Virt-manager = I mentioned this one on the virt-tools list already. Virt-manager describes a routed network as a NAT network. = Bridge = Another issue was an error while using virsh iface-define. Given br0.xml: [root at tofu2 network-scripts]# virsh iface-define br0.xml error: Failed to define interface from br0.xml error: invalid argument in virGetInterface It did work however. [root at tofu2 network-scripts]# cat ifcfg-eth0 DEVICE=eth0 HWADDR=00:24:E8:30:20:E7 ONBOOT=yes BRIDGE=br0 TYPE=Ethernet BOOTPROTO=dhcp PEERDNS=yes PEERROUTES=yes NAME="System eth0" UUID=5fb06bd0-0bb0-7ffb-45f1-d6edd65f3e03 [root at tofu2 network-scripts]# cat ifcfg-br0 DEVICE=br0 ONBOOT=yes TYPE=Bridge BOOTPROTO=none IPADDR=10.10.10.158 NETMASK=255.255.255.0 GATEWAY=10.10.10.254 But now eth0 is missing from iface-list [root at tofu2 network-scripts]# virsh iface-list Name State MAC Address -------------------------------------------- lo active 00:00:00:00:00:00 Presumably, that's because from the libvirt / netcf perspective (which is just looking at the ifcfg files) I have replaced it with the br0 interface. However, br0 is not started, so presumably that's why it isn't listed either. After a 'service network restart' things look sane again. [root at tofu2 network-scripts]# virsh iface-list Name State MAC Address -------------------------------------------- br0 active 00:24:e8:30:20:e7 lo active 00:00:00:00:00:00 = Network = And finally, after creating this bridge, it would be nice to create a network which utilizes it as follows. That fails the way I attempted to do it. [root at tofu2 ~]# cat net-bridged.xml bridged [root at tofu2 ~]# virsh net-define net-bridged.xml error: Failed to define network from net-bridged.xml error: internal error Forwarding requested, but no IPv4 address/netmask provided Presumably that's the goal of http://fedoraproject.org/wiki/Features/Shared_Network_Interface in F13. Thanks for any comments. From crobinso at redhat.com Mon Nov 16 13:05:26 2009 From: crobinso at redhat.com (Cole Robinson) Date: Mon, 16 Nov 2009 08:05:26 -0500 Subject: [fedora-virt] bridged guest networking in f12 with netcf In-Reply-To: <1205932802.823231258319422164.JavaMail.root@zebra.lib.ucdavis.edu> References: <1205932802.823231258319422164.JavaMail.root@zebra.lib.ucdavis.edu> Message-ID: <4B014E16.1040103@redhat.com> On 11/15/2009 04:10 PM, Dale Bewley wrote: > Hello all, > > I started a blog post about experimenting with netcf and configuring a guest to share the same subnet as the host. It turned into something a bit longer and more rambling of course. > > * http://tofu.org/drupal/node/86 > > At any rate I ran into a few issues while using: > virt-manager-0.8.0-7.fc12.noarch > netcf-0.1.4-1.fc12.x86_64 > libvirt-0.7.1-15.fc12.x86_64 > > = Virt-manager = > I mentioned this one on the virt-tools list already. Virt-manager describes a routed network as a NAT network. > Fixed now, as mentioned on virt-tools. > = Bridge = > Another issue was an error while using virsh iface-define. > > Given br0.xml: > > > > > > > > > > > > > > [root at tofu2 network-scripts]# virsh iface-define br0.xml > error: Failed to define interface from br0.xml > error: invalid argument in virGetInterface > > It did work however. > There were several bug fixes in this code in libvirt recently, so hopefully a later version will address that error. > [root at tofu2 network-scripts]# cat ifcfg-eth0 > DEVICE=eth0 > HWADDR=00:24:E8:30:20:E7 > ONBOOT=yes > BRIDGE=br0 > TYPE=Ethernet > BOOTPROTO=dhcp > PEERDNS=yes > PEERROUTES=yes > NAME="System eth0" > UUID=5fb06bd0-0bb0-7ffb-45f1-d6edd65f3e03 > > [root at tofu2 network-scripts]# cat ifcfg-br0 > DEVICE=br0 > ONBOOT=yes > TYPE=Bridge > BOOTPROTO=none > IPADDR=10.10.10.158 > NETMASK=255.255.255.0 > GATEWAY=10.10.10.254 > > But now eth0 is missing from iface-list > > [root at tofu2 network-scripts]# virsh iface-list > Name State MAC Address > -------------------------------------------- > lo active 00:00:00:00:00:00 > > Presumably, that's because from the libvirt / netcf perspective (which is just looking at the ifcfg files) I have replaced it with the br0 interface. However, br0 is not started, so presumably that's why it isn't listed either. After a 'service network restart' things look sane again. > > [root at tofu2 network-scripts]# virsh iface-list > Name State MAC Address > -------------------------------------------- > br0 active 00:24:e8:30:20:e7 > lo active 00:00:00:00:00:00 > > Probably want to try iface-list --all here, should list inactive interfaces. > = Network = > And finally, after creating this bridge, it would be nice to create a network which utilizes it as follows. That fails the way I attempted to do it. > > [root at tofu2 ~]# cat net-bridged.xml > > bridged > > > > Physical device bridges are referenced directly in the VM xml, not via the virtual network XML. Example interface device block to add to your VM: - Cole From dlbewley at lib.ucdavis.edu Mon Nov 16 18:51:52 2009 From: dlbewley at lib.ucdavis.edu (Dale Bewley) Date: Mon, 16 Nov 2009 10:51:52 -0800 (PST) Subject: [fedora-virt] bridged guest networking in f12 with netcf In-Reply-To: <858537169.842451258396781206.JavaMail.root@zebra.lib.ucdavis.edu> Message-ID: <278732826.843411258397512159.JavaMail.root@zebra.lib.ucdavis.edu> ----- "Cole Robinson" wrote: > On 11/15/2009 04:10 PM, Dale Bewley wrote: ... > > [root at tofu2 network-scripts]# virsh iface-define br0.xml > > error: Failed to define interface from br0.xml > > error: invalid argument in virGetInterface > > > > It did work however. ... > > But now eth0 is missing from iface-list > > > > [root at tofu2 network-scripts]# virsh iface-list > > Name State MAC Address > > -------------------------------------------- > > lo active 00:00:00:00:00:00 > > > > Presumably, that's because from the libvirt / netcf perspective > (which is just looking at the ifcfg files) I have replaced it with the > br0 interface. However, br0 is not started, so presumably that's why > it isn't listed either. After a 'service network restart' things look > sane again. > > > > [root at tofu2 network-scripts]# virsh iface-list > > Name State MAC Address > > -------------------------------------------- > > br0 active 00:24:e8:30:20:e7 > > lo active 00:00:00:00:00:00 > > Probably want to try iface-list --all here, should list inactive > interfaces. That's what I would expect, but I recreated the situation and tried that. With and without --all, only lo showed up until a network restart. > > = Network = > > And finally, after creating this bridge, it would be nice to create > a network which utilizes it as follows. That fails the way I attempted > to do it. > > > > [root at tofu2 ~]# cat net-bridged.xml > > > > bridged > > > > > > > > > > Physical device bridges are referenced directly in the VM xml, not via > the > virtual network XML. Example interface device block to add to your > VM: > > > > > > > - Cole I understand that, but I did lose track of that in my mind. It doesn't make much sense to refer to it as a 'network'. I do see that br0 shows up in the network/interface selection dialog of the VM creation wizard. One thing I just noticed is that NetworkManager is taking ownership of eth0 and preventing the br0 from being created. One fix is to chkconfig NetworkManager off chkconfig network on service NetworkManager stop service network restart I suppose the other solution would be to educate NetworkManager about the new configuration, but I've never really used paid much attention to NM. I'd have to look into how to do that properly. -------------- next part -------------- A non-text attachment was scrubbed... Name: net-select.png Type: image/png Size: 43109 bytes Desc: not available URL: From dlbewley at lib.ucdavis.edu Mon Nov 16 18:58:15 2009 From: dlbewley at lib.ucdavis.edu (Dale Bewley) Date: Mon, 16 Nov 2009 10:58:15 -0800 (PST) Subject: [fedora-virt] bridged guest networking in f12 with netcf In-Reply-To: <278732826.843411258397512159.JavaMail.root@zebra.lib.ucdavis.edu> Message-ID: <1561816066.844001258397895826.JavaMail.root@zebra.lib.ucdavis.edu> Ooops. Sorry, that was the wrong attachment which lacked br0. It does show up as expected. It wasn't that illuminating anyway, I won't bother sending the good screenshot. From crobinso at redhat.com Mon Nov 16 19:39:01 2009 From: crobinso at redhat.com (Cole Robinson) Date: Mon, 16 Nov 2009 14:39:01 -0500 Subject: [fedora-virt] bridged guest networking in f12 with netcf In-Reply-To: <278732826.843411258397512159.JavaMail.root@zebra.lib.ucdavis.edu> References: <278732826.843411258397512159.JavaMail.root@zebra.lib.ucdavis.edu> Message-ID: <4B01AA55.7060203@redhat.com> On 11/16/2009 01:51 PM, Dale Bewley wrote: > ----- "Cole Robinson" wrote: >>> = Network = >>> And finally, after creating this bridge, it would be nice to create >> a network which utilizes it as follows. That fails the way I attempted >> to do it. >>> >>> [root at tofu2 ~]# cat net-bridged.xml >>> >>> bridged >>> >>> >>> >>> >> >> Physical device bridges are referenced directly in the VM xml, not via >> the >> virtual network XML. Example interface device block to add to your >> VM: >> >> >> >> >> >> >> - Cole > > I understand that, but I did lose track of that in my mind. It doesn't > make much sense to refer to it as a 'network'. I do see that br0 shows > up in the network/interface selection dialog of the VM creation wizard. > > One thing I just noticed is that NetworkManager is taking ownership > of eth0 and preventing the br0 from being created. One fix is to > > chkconfig NetworkManager off > chkconfig network on > service NetworkManager stop > service network restart > > I suppose the other solution would be to educate NetworkManager > about the new configuration, but I've never really used paid much > attention to NM. I'd have to look into how to do that properly. > Yes, this NetworkManager pain is a long known deficiency. NM and bridges just don't play well together. - Cole From jforbes at redhat.com Wed Nov 18 21:56:30 2009 From: jforbes at redhat.com (Justin M. Forbes) Date: Wed, 18 Nov 2009 15:56:30 -0600 Subject: [fedora-virt] ANNOUNCE: Rawhide virt repo for F12 users Message-ID: <1258581390.17468.145.camel@localhost> As was done for Fedora 11 users, the tradition continues, only the locations have changed. We've set up a repository for people running Fedora 12 who would like to test the rawhide/F13 virt packages. To use it, do e.g. $> cat > /etc/yum.repos.d/fedora-virt-preview.repo << EOF [rawvirt] name=Virtualization Rawhide for Fedora 12 baseurl=http://jforbes.fedorapeople.org/virt-preview/f12/$basearch/ enabled=1 gpgcheck=0 EOF $> yum update At the moment, it contains the F-13 versions of libvirt and virt-manager, but as F-13 development continues, it will contain more. I'll send periodic mails to the list detailing the latest updates. A couple of improvements have been made this time around. Namely packages are build with mock instead of koji so that new packages can be used a BuildRequires for other new packages. Also new builds are triggered by successful koji builds of tracked packages against dist-rawhide, so the process is a bit more automated. Also, this is still a work-in-progress. The TODO list includes: - include debuginfo packages in the repo (need more quota) - find a better location than jforbes.fedorapeople.org Comments most welcome. Help with the TODO list is even more welcome :-) Thanks, Justin M. Forbes From lutter at redhat.com Thu Nov 19 18:08:43 2009 From: lutter at redhat.com (David Lutterkort) Date: Thu, 19 Nov 2009 18:08:43 +0000 Subject: [fedora-virt] bridged guest networking in f12 with netcf In-Reply-To: <278732826.843411258397512159.JavaMail.root@zebra.lib.ucdavis.edu> References: <278732826.843411258397512159.JavaMail.root@zebra.lib.ucdavis.edu> Message-ID: <1258654123.22706.0.camel@avon.watzmann.net> On Mon, 2009-11-16 at 10:51 -0800, Dale Bewley wrote: > One thing I just noticed is that NetworkManager is taking ownership > of eth0 and preventing the br0 from being created. One fix is to > > chkconfig NetworkManager off > chkconfig network on > service NetworkManager stop > service network restart > > I suppose the other solution would be to educate NetworkManager > about the new configuration That's the long-term plan, but for now, your 'fix' above is as good as it gets. David From gianluca.cecchi at gmail.com Fri Nov 20 10:20:03 2009 From: gianluca.cecchi at gmail.com (Gianluca Cecchi) Date: Fri, 20 Nov 2009 11:20:03 +0100 Subject: [fedora-virt] f11 rawvirt to f12: problems? Message-ID: <561c252c0911200220q163f805etceafcca3bf111f41@mail.gmail.com> Having F11 + fedora-virt-preview repo, with some Linux guests running without problems. Then updated to F12 After some pains with preupgrade and upgrade, finally I was on F12 Actually I probably have to prune the duplicates because: [root@ ~]# rpm -qa libvirt\* qemu\* virt-manager\* qemu-common-0.11.0-11.fc11.x86_64 virt-manager-0.8.0-7.fc12.noarch libvirt-client-0.7.1-15.fc12.x86_64 qemu-common-0.11.0-11.fc12.x86_64 qemu-img-0.11.0-11.fc11.x86_64 qemu-img-0.11.0-11.fc12.x86_64 qemu-kvm-0.11.0-11.fc12.x86_64 libvirt-python-0.7.1-15.fc12.x86_64 qemu-system-x86-0.11.0-11.fc12.x86_64 libvirt-0.7.1-15.fc12.x86_64 (in the mean time I disabled fedora-virt-preview repo) So on 18th of November I booted the server with Nov 18 17:30:12 virtfed kernel: Linux version 2.6.31.5-127.fc12.x86_64 (mockbuild at x86-4.fedora.phx.redhat.com) (gcc version 4.4.2 20091027 (Red Hat 4.4.2-7) (GCC) ) #1 SMP Sat Nov 7 21:11:14 EST 2009 . I found this trace at boot: Nov 18 17:30:12 virtfed kernel: ------------[ cut here ]------------ Nov 18 17:30:12 virtfed kernel: WARNING: at arch/x86/kernel/cpu/mtrr/generic.c:456 generic_get_mtrr+0xcc/0x10a() (Not tainted) Nov 18 17:30:12 virtfed kernel: Hardware name: ProLiant BL480c G1 Nov 18 17:30:12 virtfed kernel: mtrr: your BIOS has set up an incorrect mask, fixing it up. Nov 18 17:30:12 virtfed kernel: Modules linked in: radeon(+) ttm drm_kms_helper drm i2c_algo_bit i2c_core Nov 18 17:30:12 virtfed kernel: Pid: 183, comm: work_for_cpu Not tainted 2.6.31.5-127.fc12.x86_64 #1 Nov 18 17:30:12 virtfed kernel: Call Trace: Nov 18 17:30:12 virtfed kernel: [] warn_slowpath_common+0x84/0x9c Nov 18 17:30:12 virtfed kernel: [] warn_slowpath_fmt+0x41/0x43 Nov 18 17:30:12 virtfed kernel: [] ? bus_find_device+0x88/0x98 Nov 18 17:30:12 virtfed kernel: [] generic_get_mtrr+0xcc/0x10a Nov 18 17:30:12 virtfed kernel: [] mtrr_add_page+0x16c/0x34a Nov 18 17:30:12 virtfed kernel: [] mtrr_add+0x48/0x54 Nov 18 17:30:12 virtfed kernel: [] radeon_object_init+0x2e/0x87 [radeon] Nov 18 17:30:12 virtfed kernel: [] r100_init+0x19d/0x23c [radeon] Nov 18 17:30:12 virtfed kernel: [] ? vga_client_register+0x72/0x7d Nov 18 17:30:12 virtfed kernel: [] radeon_device_init+0x203/0x27d [radeon] Nov 18 17:30:12 virtfed kernel: [] radeon_driver_load_kms+0xff/0x13a [radeon] Nov 18 17:30:12 virtfed kernel: [] drm_get_dev+0x36e/0x46f [drm] Nov 18 17:30:12 virtfed kernel: [] ? do_work_for_cpu+0x0/0x2a Nov 18 17:30:12 virtfed kernel: [] radeon_pci_probe+0x15/0x269 [radeon] Nov 18 17:30:12 virtfed kernel: [] local_pci_probe+0x17/0x1b Nov 18 17:30:12 virtfed kernel: [] do_work_for_cpu+0x18/0x2a Nov 18 17:30:12 virtfed kernel: [] kthread+0x91/0x99 Nov 18 17:30:12 virtfed kernel: [] child_rip+0xa/0x20 Nov 18 17:30:12 virtfed kernel: [] ? kthread+0x0/0x99 Nov 18 17:30:12 virtfed kernel: [] ? child_rip+0x0/0x20 Nov 18 17:30:12 virtfed kernel: ---[ end trace 5173c14fe23b4391 ]--- the system is a standalone node of a cluster (based on cman/rgmanager as provided in F11, now F12) I started two CentOS 5.4 vms and started ksm and ksmtuned [root at virtfed ~]# cat /sys/kernel/mm/ksm/pages_shared 2000 This morning I found the two guests in shutdown mode and these messages: Nov 20 08:22:06 virtfed kernel: script.sh invoked oom-killer: gfp_mask=0x201da, order=0, oomkilladj=0 Nov 20 08:22:06 virtfed kernel: script.sh cpuset=/ mems_allowed=0 Nov 20 08:22:06 virtfed kernel: Pid: 14360, comm: script.sh Tainted: G W 2.6.31.5-127.fc12.x86_64 #1 Nov 20 08:22:06 virtfed kernel: Call Trace: Nov 20 08:22:06 virtfed kernel: [] ? cpuset_print_task_mems_allowed+0x91/0x9d Nov 20 08:22:06 virtfed kernel: [] oom_kill_process+0x98/0x256 Nov 20 08:22:06 virtfed kernel: [] ? select_bad_process+0xa3/0x102 Nov 20 08:22:06 virtfed kernel: [] __out_of_memory+0x8a/0x99 Nov 20 08:22:06 virtfed kernel: [] out_of_memory+0x163/0x195 Nov 20 08:22:06 virtfed kernel: [] __alloc_pages_nodemask+0x491/0x584 Nov 20 08:22:06 virtfed kernel: [] alloc_pages_current+0x95/0x9e Nov 20 08:22:06 virtfed kernel: [] __page_cache_alloc+0x5f/0x61 Nov 20 08:22:06 virtfed kernel: [] __do_page_cache_readahead+0x98/0x176 Nov 20 08:22:06 virtfed kernel: [] ra_submit+0x21/0x25 Nov 20 08:22:06 virtfed kernel: [] filemap_fault+0x193/0x317 Nov 20 08:22:06 virtfed kernel: [] __do_fault+0x54/0x3c4 Nov 20 08:22:06 virtfed kernel: [] handle_mm_fault+0x2f6/0x705 Nov 20 08:22:06 virtfed kernel: [] ? virt_to_head_page+0xe/0x2f Nov 20 08:22:06 virtfed kernel: [] ? free_bprm+0x44/0x49 Nov 20 08:22:06 virtfed kernel: [] do_page_fault+0x281/0x299 Nov 20 08:22:06 virtfed kernel: [] page_fault+0x25/0x30 Nov 20 08:22:06 virtfed kernel: Mem-Info: Nov 20 08:22:06 virtfed kernel: Node 0 DMA per-cpu: Nov 20 08:22:06 virtfed kernel: CPU 0: hi: 0, btch: 1 usd: 0 Nov 20 08:22:06 virtfed kernel: CPU 1: hi: 0, btch: 1 usd: 0 Nov 20 08:22:06 virtfed kernel: CPU 2: hi: 0, btch: 1 usd: 0 Nov 20 08:22:06 virtfed kernel: CPU 3: hi: 0, btch: 1 usd: 0 Nov 20 08:22:06 virtfed kernel: CPU 4: hi: 0, btch: 1 usd: 0 Nov 20 08:22:06 virtfed kernel: CPU 5: hi: 0, btch: 1 usd: 0 Nov 20 08:22:06 virtfed kernel: CPU 6: hi: 0, btch: 1 usd: 0 Nov 20 08:22:06 virtfed kernel: CPU 7: hi: 0, btch: 1 usd: 0 Nov 20 08:22:06 virtfed kernel: Node 0 DMA32 per-cpu: Nov 20 08:22:06 virtfed kernel: CPU 0: hi: 186, btch: 31 usd: 164 Nov 20 08:22:06 virtfed kernel: CPU 1: hi: 186, btch: 31 usd: 23 Nov 20 08:22:06 virtfed kernel: CPU 2: hi: 186, btch: 31 usd: 100 Nov 20 08:22:06 virtfed kernel: CPU 3: hi: 186, btch: 31 usd: 0 Nov 20 08:22:06 virtfed kernel: CPU 4: hi: 186, btch: 31 usd: 47 Nov 20 08:22:06 virtfed kernel: CPU 5: hi: 186, btch: 31 usd: 51 Nov 20 08:22:06 virtfed kernel: CPU 6: hi: 186, btch: 31 usd: 181 Nov 20 08:22:06 virtfed kernel: CPU 7: hi: 186, btch: 31 usd: 48 Nov 20 08:22:06 virtfed kernel: Node 0 Normal per-cpu: Nov 20 08:22:06 virtfed kernel: CPU 0: hi: 186, btch: 31 usd: 160 Nov 20 08:22:06 virtfed kernel: CPU 1: hi: 186, btch: 31 usd: 5 Nov 20 08:22:06 virtfed kernel: CPU 2: hi: 186, btch: 31 usd: 181 Nov 20 08:22:06 virtfed kernel: CPU 3: hi: 186, btch: 31 usd: 0 Nov 20 08:22:06 virtfed kernel: CPU 4: hi: 186, btch: 31 usd: 73 Nov 20 08:22:06 virtfed kernel: CPU 5: hi: 186, btch: 31 usd: 38 Nov 20 08:22:06 virtfed kernel: CPU 6: hi: 186, btch: 31 usd: 156 Nov 20 08:22:06 virtfed kernel: CPU 7: hi: 186, btch: 31 usd: 31 Nov 20 08:22:06 virtfed kernel: Active_anon:2563037 active_file:0 inactive_anon:368313 Nov 20 08:22:06 virtfed kernel: inactive_file:31 unevictable:12620 dirty:0 writeback:0 unstable:0 Nov 20 08:22:06 virtfed kernel: free:16269 slab:25512 mapped:7756 pagetables:10201 bounce:0 Nov 20 08:22:06 virtfed kernel: Node 0 DMA free:15824kB min:16kB low:20kB high:24kB active_anon:0kB inactive_anon:0kB active_file:0kB inactive_file:0kB unevictable:0kB present:15320kB pages_scanned:0 all_unreclaimable? yes Nov 20 08:22:06 virtfed kernel: lowmem_reserve[]: 0 3254 12092 12092 Nov 20 08:22:06 virtfed kernel: Node 0 DMA32 free:39040kB min:3784kB low:4728kB high:5676kB active_anon:2557420kB inactive_anon:511296kB active_file:0kB inactive_file:0kB unevictable:804kB present:3332660kB pages_scanned:0 all_unreclaimable? no Nov 20 08:22:06 virtfed kernel: lowmem_reserve[]: 0 0 8837 8837 Nov 20 08:22:06 virtfed kernel: Node 0 Normal free:10212kB min:10280kB low:12848kB high:15420kB active_anon:7694728kB inactive_anon:961956kB active_file:0kB inactive_file:236kB unevictable:49676kB present:9049596kB pages_scanned:0 all_unreclaimable? no Nov 20 08:22:06 virtfed kernel: lowmem_reserve[]: 0 0 0 0 Nov 20 08:22:06 virtfed kernel: Node 0 DMA: 2*4kB 1*8kB 2*16kB 1*32kB 2*64kB 0*128kB 1*256kB 0*512kB 1*1024kB 1*2048kB 3*4096kB = 15824kB Nov 20 08:22:06 virtfed kernel: Node 0 DMA32: 1068*4kB 1263*8kB 511*16kB 161*32kB 25*64kB 2*128kB 1*256kB 6*512kB 0*1024kB 1*2048kB 1*4096kB = 39032kB Nov 20 08:22:06 virtfed kernel: Node 0 Normal: 1595*4kB 0*8kB 1*16kB 1*32kB 1*64kB 1*128kB 1*256kB 1*512kB 1*1024kB 1*2048kB 0*4096kB = 10460kB Nov 20 08:22:06 virtfed kernel: 25465 total pagecache pages Nov 20 08:22:06 virtfed kernel: 17663 pages in swap cache Nov 20 08:22:06 virtfed kernel: Swap cache stats: add 1210138, delete 1192475, find 119729/133852 Nov 20 08:22:06 virtfed kernel: Free swap = 0kB Nov 20 08:22:06 virtfed kernel: Total swap = 4194296kB Nov 20 08:22:06 virtfed kernel: 3145727 pages RAM Nov 20 08:22:06 virtfed kernel: 63020 pages reserved Nov 20 08:22:06 virtfed kernel: 32665 pages shared Nov 20 08:22:06 virtfed kernel: 3054335 pages non-shared Nov 20 08:22:06 virtfed kernel: Out of memory: kill process 14357 (qemu-kvm) score 994666 or a child Nov 20 08:22:06 virtfed kernel: Killed process 14357 (qemu-kvm) Nov 20 08:22:06 virtfed kernel: virbr0: port 2(vnet1) entering disabled state Nov 20 08:22:06 virtfed kernel: device vnet1 left promiscuous mode Nov 20 08:22:06 virtfed kernel: virbr0: port 2(vnet1) entering disabled state and then similar about 1h 20minutes after for the other one.... Now I have: [root at virtfed ~]# free total used free shared buffers cached Mem: 12330828 1866848 10463980 0 546088 158820 -/+ buffers/cache: 1161940 11168888 Swap: 4194296 30024 4164272 and [root@ ~]# cat /sys/kernel/mm/ksm/pages_shared 0 It seems script.sh is the one provided by cluster manager but following messages were about other processed Nov 20 09:45:25 virtfed kernel: ksmd invoked oom-killer: gfp_mask=0x200d2, order=0, oomkilladj=0 Nov 20 09:45:25 virtfed kernel: ksmd cpuset=/ mems_allowed=0 Nov 20 10:58:35 virtfed kernel: awk invoked oom-killer: gfp_mask=0x201da, order=0, oomkilladj=0 Nov 20 10:58:35 virtfed kernel: awk cpuset=/ mems_allowed=0 Nov 20 10:58:35 virtfed kernel: libvirtd invoked oom-killer: gfp_mask=0x201da, order=0, oomkilladj=0 Nov 20 10:58:35 virtfed kernel: libvirtd cpuset=/ mems_allowed=0 /etc/sysconfig/ksm is empty (apart comments) at this moment: [root at virtfed log]# service ksm status ksm is not running [root at virtfed log]# service ksmtuned status ksmtuned (pid 8672) is running... Any hints? Thanks, Gianluca From kaboom at oobleck.net Fri Nov 20 15:28:58 2009 From: kaboom at oobleck.net (Chris Ricker) Date: Fri, 20 Nov 2009 10:28:58 -0500 (EST) Subject: [fedora-virt] Sound status on VM? In-Reply-To: References: Message-ID: On Mon, 9 Nov 2009, Paul Lambert wrote: > Does sound now work with a VM?? I am using rawhide x86_64 for the host and > FE-11 x86 for the guest.? I also have a Windows XP VM.? The VM hardware > shows es1370.? The host sound plays mp3s using either Totem or Rhythmbox > just fine but no sound from either VM. For the Windows guest, you can try forwarding sound through RDP: padsp rdesktop -r sound:local later, chris From markmc at redhat.com Fri Nov 20 20:02:41 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Fri, 20 Nov 2009 20:02:41 +0000 Subject: [fedora-virt] Fedora virt status Message-ID: <1258747361.10368.5.camel@blaa> Fedora 12 Release ================= Yay! Fedora 12 was released on Tuesday: http://fedoraproject.org/wiki/Fedora_12_Announcement Thanks to all who help get this release out! There was a single response to my call for a "witty tagline" for Fedora 12, so I went with Avi's "Dirty Dozen" suggestion - it's Fedora *12* and we have *12* virt features listed so ... :-) https://fedoraproject.org/wiki/Virtualization/History#Fedora_12:_The_Dirty_Dozen Note also that the marketing folks have posted an interview with some of the Fedora virt developers here: https://fedoraproject.org/wiki/Virtualization_improvements_in_Fedora_12 Fedora 12 includes a number of improvements in the field of Virtualization. New tools enable system administrators to perform nearly impossible - until now - tasks easily. Imagine re-configuring a virtual machine off-line, add new hardware to VM with out restarting it, migrate to another host without restarting the VMs and many other exotic features. Let's hear what developers have to say about those wonderful new options. virt-preview ============ Justin Forbes announced the availability of a virt-preview repository for Fedora 12 users: http://www.redhat.com/archives/fedora-virt/2009-November/msg00041.html A couple of improvements have been made this time around. Namely packages are build with mock instead of koji so that new packages can be used a BuildRequires for other new packages. Also new builds are triggered by successful koji builds of tracked packages against dist-rawhide, so the process is a bit more automated. These improvements should make virt-preview a hell of a lot more useful and manageable. Kudos Justin! Release Blockers ================ As with any release, we went through a couple of fire-drills with last-minute serious blocker bugs: https://bugzilla.redhat.com/526549 rawhide/i386 kvm host corrupts data of guests https://bugzilla.redhat.com/533063 preadv()/pwritev() prototypes are broken on i386 with -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 KVM guests on 32 bit hosts were seeing data corruption in some circumstances. After huge efforts from Justin Forbes, James Laska and Milan Broz to get the issue narrowed down to qemu, yours truly bisected it to the introduction of preadv/pwritev support. I happened to be sitting around a table with Dan Berrange, Chris Wright, Stephen Tweedie, Herbert Xu and Dor Laor at the time and all chipped in helping to figure out the issue. Eventually, danpb found the issue with glibc's preadv()/pwritev() prototypes in its headers and a simple fix sorted the issue out just in time for the release. Now there's a team effort! Another saga was going on in parallel on the VT-d front: https://bugzilla.redhat.com/524808 swiotlb should be enabled when VT-d setup fails This got recognized as a critical issue very late in the release as it was realized that there are plenty of users who will have VT-d enabled in the BIOS, certain broken BIOSes and >4Gb memory (including swap). In the end dwmw2's patch was included and tested by multiple people, so intel_iommu=on is still the default in Fedora 12. https://bugzilla.redhat.com/533952 DMAR: kernel panic with 2.6.31.5-127.fc12.i686 But then this issue came to prominence post-release, and it turns out the fix for bug #524808 broke those machines with fault BIOSes and <4Gb of RAM. Sometimes you just can't win! Lots of folks have been battling away at these bugs including David Woodhouse, Chris Wright, Adam Williamson and a whole bunch of others. Mount guest filesystems in the host =================================== Rich Jones continues rocking with yet another cool feature: http://www.redhat.com/archives/fedora-virt/2009-November/msg00001.html We just built a package called 'libguestfs-mount' in Rawhide which lets you mount virtual machines' filesystems on the host, using FUSE. Rich may have announced this very quietly, but sit up and take note - this is damn useful! Bugs ==== DOOM-O-METER: 186 bugs open three weeks ago, only 173 now! == Ongoing == === kernel === https://bugzilla.redhat.com/532009 "kvm_run: failed entry, reason 7" when running guest with F11 LiveCD Looks like an issue starting a guest on a F11 AMD host. === qemu ==== https://bugzilla.redhat.com/539365 qemu's system_powerdown doesn't work with Windows XP Now that virtinst enables acpi for XP guests, we're noticing that 'virsh shutdown' has no effect on them. That may be just due to a lack of ACPI support in the guest. https://bugzilla.redhat.com/id=538047 Any use of QEMU's vvfat driver always abort()s the whole process Dan Berrange discovered this nasty hack which causes glibc to abort qemu. === libvirt === https://bugzilla.redhat.com/537938 restore fails with large files; libvirtd becomes unresponsive Charles Duffy reported this issue and proposed a fix. Apparently this isn't needed upstream anymore, but we may still need it for F-12. https://bugzilla.redhat.com/532654 virsh save hangs in Fedora12 rawhide SELinux related issues breaking save/restore. https://bugzilla.redhat.com/536760 SELinux is preventing /usr/bin/qemu-kvm "write" access on sr0 It looks like we're getting AVCs because qemu is trying to open read-only images for writing. === python-virtinst === https://bugzilla.redhat.com/492082 python-virtinst: add a label to newly created disks? This has come back full circle again. Now that the anaconda guys have refused to do anything about it, our only option is to figure out some way of labelling the disk before starting the guest. === virt-manager === https://bugzilla.redhat.com/532216 virt-manager should disable unimplemented reboot option for qemu/kvm libvirt doesn't implement reboot for KVM (#496537), yet the button is still available in virt-manager. https://bugzilla.redhat.com/537221 virt-manager: URL install should prompt to change scratchdir perms virt-manager is downloading kernel and initrds to the users home directory where qemu cannot read them. Proposal is that virt-manager should prompt to change the ACLs on this dir like it does for ISOs. == Resolved == === kernel === https://bugzilla.redhat.com/532215 KSM bad_page() issue preventing VM startup Justin Forbes has cherry-picked a fix from upsteam into the kernel in updates. === qemu === https://bugzilla.redhat.com/539583 qemu use-after-free crash in slirp/m_free() A couple of days after the F-12 release, we get a report of a very obvious crasher in qemu 0.11.0's slirp code. One really does wonder how these things get unnoticed until after the release :-) https://bugzilla.redhat.com/533573 qemu-img convert can't handle parallels images above 4GiB David Woodhouse sent a patch upstream to fix this. === xen === https://bugzilla.redhat.com/521800 kernel backtrace: possible recursive locking detected on Xen domU Apparently this problem just disappeared with a recent 2.6.31 kernel. From rpjday at crashcourse.ca Sat Nov 21 15:40:45 2009 From: rpjday at crashcourse.ca (Robert P. J. Day) Date: Sat, 21 Nov 2009 10:40:45 -0500 (EST) Subject: [fedora-virt] release notes virt passage seems overly pessimistic Message-ID: reading here: http://docs.fedoraproject.org/release-notes/f12/en-US/html/sect-Release_Notes-Virtualization.html down at the bottom: "KVM requires hardware virtualization features in the host system. Systems lacking hardware virtualization do not support Xen guests at this time." if one is new to virt, that could be read as, "without HW virt, you're pretty much screwed." or could that be worded a bit differently? rday -- ======================================================================== Robert P. J. Day Waterloo, Ontario, CANADA Linux Consulting, Training and Kernel Pedantry. Web page: http://crashcourse.ca Twitter: http://twitter.com/rpjday ======================================================================== From bderzhavets at yahoo.com Sat Nov 21 15:49:58 2009 From: bderzhavets at yahoo.com (Boris Derzhavets) Date: Sat, 21 Nov 2009 07:49:58 -0800 (PST) Subject: [fedora-virt] release notes virt passage seems overly pessimistic In-Reply-To: Message-ID: <263099.24313.qm@web56105.mail.re3.yahoo.com> > if one is new to virt, that could be read as, "without HW virt, > you're pretty much screwed." ??? I believe the answer is yes. Boris. --- On Sat, 11/21/09, Robert P. J. Day wrote: From: Robert P. J. Day Subject: [fedora-virt] release notes virt passage seems overly pessimistic To: "Fedora Virtualization Mailing List" Date: Saturday, November 21, 2009, 10:40 AM ? reading here: http://docs.fedoraproject.org/release-notes/f12/en-US/html/sect-Release_Notes-Virtualization.html down at the bottom: "KVM requires hardware virtualization features in the host system. Systems lacking hardware virtualization do not support Xen guests at this time." ? if one is new to virt, that could be read as, "without HW virt, you're pretty much screwed."? or could that be worded a bit differently? rday -- ======================================================================== Robert P. J. Day? ? ? ? ? ? ? ? ? ? ? ? ? ? ???Waterloo, Ontario, CANADA ? ? ? ? ? ? Linux Consulting, Training and Kernel Pedantry. Web page:? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? http://crashcourse.ca Twitter:? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ???http://twitter.com/rpjday ======================================================================== _______________________________________________ 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 psj at familyjenner.co.uk Sat Nov 21 20:43:13 2009 From: psj at familyjenner.co.uk (Paul Jenner) Date: Sat, 21 Nov 2009 20:43:13 +0000 Subject: [fedora-virt] ANNOUNCE: Rawhide virt repo for F12 users In-Reply-To: <1258581390.17468.145.camel@localhost> References: <1258581390.17468.145.camel@localhost> Message-ID: <1258836193.9361.3.camel@82-71-88-77.dsl.in-addr.zen.co.uk> Hi Justin. On Wed, 2009-11-18 at 15:56 -0600, Justin M. Forbes wrote: > As was done for Fedora 11 users, the tradition continues, only the > locations have changed. > > We've set up a repository for people running Fedora 12 who would like > to test the rawhide/F13 virt packages. To use it, do e.g. > > $> cat > /etc/yum.repos.d/fedora-virt-preview.repo << EOF > [rawvirt] > name=Virtualization Rawhide for Fedora 12 > baseurl=http://jforbes.fedorapeople.org/virt-preview/f12/$basearch/ > enabled=1 > gpgcheck=0 > EOF > $> yum update I guess you have this on the list but the wiki page should be updated to match the new repo location: http://fedoraproject.org/wiki/Virtualization_Preview_Repository It is referenced from the F12 release note - Section 5.3.12.2. Virtualization Technology Preview Repo - but still points users at the F11 repo. I would do it myself but I don't have a wiki login (poor excuse for not getting one I know :-)) Paul -- Paul Jenner From dlbewley at lib.ucdavis.edu Sun Nov 22 17:29:47 2009 From: dlbewley at lib.ucdavis.edu (Dale Bewley) Date: Sun, 22 Nov 2009 09:29:47 -0800 (PST) Subject: [fedora-virt] release notes virt passage seems overly pessimistic In-Reply-To: <14311526.211258910943940.JavaMail.dlbewley@seitan.home.bewley.net> Message-ID: <26210310.231258910982840.JavaMail.dlbewley@seitan.home.bewley.net> ----- "Robert P. J. Day" wrote: > reading here: > > http://docs.fedoraproject.org/release-notes/f12/en-US/html/sect-Release_Notes-Virtualization.html > > down at the bottom: > > "KVM requires hardware virtualization features in the host system. > > Systems lacking hardware virtualization do not support Xen guests at > this time." > > if one is new to virt, that could be read as, "without HW virt, > you're pretty much screwed." or could that be worded a bit > differently? This is the same text carried over from 10 and 11. http://docs.fedoraproject.org/release-notes/f10/en_US/What_Do_System_Adminstrators_Care_About.html#sn-Virtualization While Xen does not require hardware support, KVM still (always will?) does. Xen guests on Fedora still require KVM-based xenner. There has been success with Xen dom0 on experimental 3rd party kernels, but that may be beyond the scope of the release notes. We wouldn't want it to be inferred as a recommendation. Hopefully the F13 release notes will be able to describe native support for Xen dom0 hosts. http://fedoraproject.org/wiki/FWN/Issue202#No_Xen_dom0_in_Fedora_12_Hopefully_13 If you have any suggestions for further improvement to the release notes, please keep them coming. -- Dale From rpjday at crashcourse.ca Sun Nov 22 17:47:09 2009 From: rpjday at crashcourse.ca (Robert P. J. Day) Date: Sun, 22 Nov 2009 12:47:09 -0500 (EST) Subject: [fedora-virt] release notes virt passage seems overly pessimistic In-Reply-To: <26210310.231258910982840.JavaMail.dlbewley@seitan.home.bewley.net> References: <26210310.231258910982840.JavaMail.dlbewley@seitan.home.bewley.net> Message-ID: On Sun, 22 Nov 2009, Dale Bewley wrote: > ----- "Robert P. J. Day" wrote: > > reading here: > > > > http://docs.fedoraproject.org/release-notes/f12/en-US/html/sect-Release_Notes-Virtualization.html > > > > down at the bottom: > > > > "KVM requires hardware virtualization features in the host system. > > > > Systems lacking hardware virtualization do not support Xen guests > > at this time." > > > > if one is new to virt, that could be read as, "without HW virt, > > you're pretty much screwed." or could that be worded a bit > > differently? > > This is the same text carried over from 10 and 11. > > http://docs.fedoraproject.org/release-notes/f10/en_US/What_Do_System_Adminstrators_Care_About.html#sn-Virtualization > > While Xen does not require hardware support, KVM still (always > will?) does. Xen guests on Fedora still require KVM-based xenner. > There has been success with Xen dom0 on experimental 3rd party > kernels, but that may be beyond the scope of the release notes. We > wouldn't want it to be inferred as a recommendation. > > Hopefully the F13 release notes will be able to describe native > support for Xen dom0 hosts. > > http://fedoraproject.org/wiki/FWN/Issue202#No_Xen_dom0_in_Fedora_12_Hopefully_13 > > If you have any suggestions for further improvement to the release > notes, please keep them coming. for people new to virtualization, i was simply suggesting that that wording still leaves some doubt as to what's possible. what about a slightly longer explanation which describes what you can support based on the capabilities of your system, as in: 1) if your system supports H/W virtualization, you can do the following: ... list of things ... 2) if your system does *not* support H/W virtualization, you are limited to the following: ... much shorter list ... that should be written for the newbie since the most frustrating experience for beginners is to invest considerable time trying to do something, only to eventually learn that it wasn't possible all along. i'm thinking a page entitled something like "So, you have a computer and you want to get into virtualization." does such a page exist? rday -- ======================================================================== Robert P. J. Day Waterloo, Ontario, CANADA Linux Consulting, Training and Kernel Pedantry. Web page: http://crashcourse.ca Twitter: http://twitter.com/rpjday ======================================================================== From dlbewley at lib.ucdavis.edu Mon Nov 23 00:34:13 2009 From: dlbewley at lib.ucdavis.edu (Dale Bewley) Date: Sun, 22 Nov 2009 16:34:13 -0800 (PST) Subject: [fedora-virt] release notes virt passage seems overly pessimistic In-Reply-To: <4729478.271258936430756.JavaMail.dlbewley@seitan.home.bewley.net> Message-ID: <9221562.291258936449830.JavaMail.dlbewley@seitan.home.bewley.net> ----- "Robert P. J. Day" wrote: > On Sun, 22 Nov 2009, Dale Bewley wrote: > > ----- "Robert P. J. Day" wrote: > > > reading here: > > > > > > > http://docs.fedoraproject.org/release-notes/f12/en-US/html/sect-Release_Notes-Virtualization.html > > > > > > down at the bottom: > > > > > > "KVM requires hardware virtualization features in the host > system. > > > > > > Systems lacking hardware virtualization do not support Xen guests > > > at this time." > > > > > > if one is new to virt, that could be read as, "without HW virt, > > > you're pretty much screwed." or could that be worded a bit > > > differently? > > > > This is the same text carried over from 10 and 11. > > > > > http://docs.fedoraproject.org/release-notes/f10/en_US/What_Do_System_Adminstrators_Care_About.html#sn-Virtualization > > > > While Xen does not require hardware support, KVM still (always > > will?) does. Xen guests on Fedora still require KVM-based xenner. > > There has been success with Xen dom0 on experimental 3rd party > > kernels, but that may be beyond the scope of the release notes. We > > wouldn't want it to be inferred as a recommendation. > > > > Hopefully the F13 release notes will be able to describe native > > support for Xen dom0 hosts. > > > > > http://fedoraproject.org/wiki/FWN/Issue202#No_Xen_dom0_in_Fedora_12_Hopefully_13 > > > > If you have any suggestions for further improvement to the release > > notes, please keep them coming. > > for people new to virtualization, i was simply suggesting that that > wording still leaves some doubt as to what's possible. > > what about a slightly longer explanation which describes what you > can support based on the capabilities of your system, as in: > > 1) if your system supports H/W virtualization, you can do the > following: > > ... list of things ... > > 2) if your system does *not* support H/W virtualization, you are > limited to the following: > > ... much shorter list ... > > that should be written for the newbie since the most frustrating > experience for beginners is to invest considerable time trying to do > something, only to eventually learn that it wasn't possible all > along. > > i'm thinking a page entitled something like "So, you have a > computer > and you want to get into virtualization." does such a page exist? > > rday Under http://fedoraproject.org/wiki/Category:Virtualization We do have http://fedoraproject.org/wiki/Getting_started_with_virtualization -- Dale Bewley - Unix Administrator - Shields Library - UC Davis GPG: 0xB098A0F3 0D5A 9AEB 43F4 F84C 7EFD 1753 064D 2583 B098 A0F3 From keqin.hong at gmail.com Mon Nov 23 03:24:17 2009 From: keqin.hong at gmail.com (Keqin Hong) Date: Mon, 23 Nov 2009 11:24:17 +0800 Subject: [fedora-virt] release notes virt passage seems overly pessimistic In-Reply-To: <9221562.291258936449830.JavaMail.dlbewley@seitan.home.bewley.net> References: <4729478.271258936430756.JavaMail.dlbewley@seitan.home.bewley.net> <9221562.291258936449830.JavaMail.dlbewley@seitan.home.bewley.net> Message-ID: I agree with Robert, as I could only use VirtualBox or QEMU on my laptop without Intel-vt. But on the other hand, I think KVM will be ideal for new machines with HW virt support especially for data centers. It's fast, mainstream and becoming mature. Hope Xen dom0 could be included as an additional choice for desktop users. keqin.hong at gmail.com On Mon, Nov 23, 2009 at 8:34 AM, Dale Bewley wrote: > ----- "Robert P. J. Day" wrote: >> On Sun, 22 Nov 2009, Dale Bewley wrote: >> > ----- "Robert P. J. Day" wrote: >> > > reading here: >> > > >> > > >> http://docs.fedoraproject.org/release-notes/f12/en-US/html/sect-Release_Notes-Virtualization.html >> > > >> > > down at the bottom: >> > > >> > > "KVM requires hardware virtualization features in the host >> system. >> > > >> > > Systems lacking hardware virtualization do not support Xen guests >> > > at this time." >> > > >> > > ? if one is new to virt, that could be read as, "without HW virt, >> > > you're pretty much screwed." ?or could that be worded a bit >> > > differently? >> > >> > This is the same text carried over from 10 and 11. >> > >> > >> http://docs.fedoraproject.org/release-notes/f10/en_US/What_Do_System_Adminstrators_Care_About.html#sn-Virtualization >> > >> > While Xen does not require hardware support, KVM still (always >> > will?) does. Xen guests on Fedora still require KVM-based xenner. >> > There has been success with Xen dom0 on experimental 3rd party >> > kernels, but that may be beyond the scope of the release notes. We >> > wouldn't want it to be inferred as a recommendation. >> > >> > Hopefully the F13 release notes will be able to describe native >> > support for Xen dom0 hosts. >> > >> > >> http://fedoraproject.org/wiki/FWN/Issue202#No_Xen_dom0_in_Fedora_12_Hopefully_13 >> > >> > If you have any suggestions for further improvement to the release >> > notes, please keep them coming. >> >> ? for people new to virtualization, i was simply suggesting that that >> wording still leaves some doubt as to what's possible. >> >> ? what about a slightly longer explanation which describes what you >> can support based on the capabilities of your system, as in: >> >> ? 1) if your system supports H/W virtualization, you can do the >> following: >> >> ? ? ?... list of things ... >> >> ? 2) if your system does *not* support H/W virtualization, you are >> limited to the following: >> >> ? ? ?... much shorter list ... >> >> ? that should be written for the newbie since the most frustrating >> experience for beginners is to invest considerable time trying to do >> something, only to eventually learn that it wasn't possible all >> along. >> >> ? i'm thinking a page entitled something like "So, you have a >> computer >> and you want to get into virtualization." ?does such a page exist? >> >> rday > > Under > http://fedoraproject.org/wiki/Category:Virtualization > > We do have > http://fedoraproject.org/wiki/Getting_started_with_virtualization > > -- > 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 > From andres at verot.com Mon Nov 23 11:04:20 2009 From: andres at verot.com (=?UTF-8?B?QW5kcsOpcyBHYXJjw61h?=) Date: Mon, 23 Nov 2009 12:04:20 +0100 Subject: [fedora-virt] Host lock up in F12 In-Reply-To: <1257775967.3007.31.camel@blaa> References: <4AF1C9C5.3000609@verot.com> <1257775967.3007.31.camel@blaa> Message-ID: <4B0A6C34.60602@verot.com> Hi, I have found a way to consistently lock up a F12 host. I had three guests, another F12 and two windows ones. I added another with the lastest Mandriva x64, configured as 'kernel newer that 2.6.25 with virtio', it installed fined and everything seemed to be all right. To test what would happen I then went to a terminal a typed 'reboot'. On powerin on, the host would allway lock up, fortunatley it gave enough time to disable Mandriva starting on boot and now it works. Does that ring a bell? Is there a way have a guest boot five minutes after boot? > Best bet is probably to hook up a serial console so you can get the full > oops info > > You mean like this? http://www.howtoforge.com/setting_up_a_serial_console Problem is I don't have a serial cable, I will have to check if the electronics store has one. Thanks, Andres From markmc at redhat.com Mon Nov 23 11:07:53 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Mon, 23 Nov 2009 11:07:53 +0000 Subject: [fedora-virt] Host lock up in F12 In-Reply-To: <4B0A6C34.60602@verot.com> References: <4AF1C9C5.3000609@verot.com> <1257775967.3007.31.camel@blaa> <4B0A6C34.60602@verot.com> Message-ID: <1258974473.16952.4.camel@blaa> Hi Andr?s, On Mon, 2009-11-23 at 12:04 +0100, Andr?s Garc?a wrote: > On powerin on, the host would allway lock up, fortunatley it gave enough > time to disable > Mandriva starting on boot and now it works. > > Does that ring a bell? Yeah, it sounds similar to https://bugzilla.redhat.com/528245 Could you also try making your network connection available to all users? See comment #7 Thanks, Mark. From phil at pricom.com.au Mon Nov 23 14:54:12 2009 From: phil at pricom.com.au (Philip Rhoades) Date: Tue, 24 Nov 2009 01:54:12 +1100 Subject: [fedora-virt] Routing to guests Message-ID: <4B0AA214.9050407@pricom.com.au> People, I have an ADSL router: 192.168.0.1 a 64bit F12 host: eth0: 192.168.0.10 virbr0: 192.168.122.1 a F11 guest: eth0: 192.168.122.68 I can ssh from/to the host/guest OK but how do I set up a route (or whatever is necessary) so that another machine: eth0: 192.168.0.12 can ssh to the guest? - "ssh 192.168.122.68" gives "no route to host" - I have "route add" things with no success. - I had a look at: http://docs.fedoraproject.org/virtualization-guide/f12/en-US/html/ but the problem does not seem to be covered there. Thanks, Phil. -- Philip Rhoades GPO Box 3411 Sydney NSW 2001 Australia E-mail: phil at pricom.com.au From jforbes at redhat.com Mon Nov 23 15:47:33 2009 From: jforbes at redhat.com (Justin M. Forbes) Date: Mon, 23 Nov 2009 09:47:33 -0600 Subject: [fedora-virt] ANNOUNCE: Rawhide virt repo for F12 users In-Reply-To: <1258836193.9361.3.camel@82-71-88-77.dsl.in-addr.zen.co.uk> References: <1258581390.17468.145.camel@localhost> <1258836193.9361.3.camel@82-71-88-77.dsl.in-addr.zen.co.uk> Message-ID: <1258991253.2509.13.camel@localhost> On Sat, 2009-11-21 at 20:43 +0000, Paul Jenner wrote: > Hi Justin. > > On Wed, 2009-11-18 at 15:56 -0600, Justin M. Forbes wrote: > > As was done for Fedora 11 users, the tradition continues, only the > > locations have changed. > > > > We've set up a repository for people running Fedora 12 who would like > > to test the rawhide/F13 virt packages. To use it, do e.g. > > > > $> cat > /etc/yum.repos.d/fedora-virt-preview.repo << EOF > > [rawvirt] > > name=Virtualization Rawhide for Fedora 12 > > baseurl=http://jforbes.fedorapeople.org/virt-preview/f12/$basearch/ > > enabled=1 > > gpgcheck=0 > > EOF > > $> yum update > > I guess you have this on the list but the wiki page should be updated to > match the new repo location: > > http://fedoraproject.org/wiki/Virtualization_Preview_Repository > > It is referenced from the F12 release note - Section 5.3.12.2. > Virtualization Technology Preview Repo - but still points users at the > F11 repo. > > I would do it myself but I don't have a wiki login (poor excuse for not > getting one I know :-)) > Thanks for pointing that out. Looks like someone got to it before me, but things look correct now. Justin From andres at verot.com Mon Nov 23 17:56:41 2009 From: andres at verot.com (=?UTF-8?B?QW5kcsOpcyBHYXJjw61h?=) Date: Mon, 23 Nov 2009 18:56:41 +0100 Subject: [fedora-virt] Host lock up in F12 In-Reply-To: <1258974473.16952.4.camel@blaa> References: <4AF1C9C5.3000609@verot.com> <1257775967.3007.31.camel@blaa> <4B0A6C34.60602@verot.com> <1258974473.16952.4.camel@blaa> Message-ID: <4B0ACCD9.5050904@verot.com> Hi > Yeah, it sounds similar to https://bugzilla.redhat.com/528245 > > Could you also try making your network connection available to all > users? See comment #7 > I don't have wireless like in that report but I do have two network cards, eth0 for the host and eth1 bridged for the guests. So to try it I added NM_CONTROLLED=yes for eth0 and NM_CONTROLLED=no for eth1 and br1, activated NetworkManager back and made sure that eth0 was available to all users. Booting Mandriva kept locking the host, so I created another mandriva virtual machine except this time I configured it as 'kernel 2.6.x', no virtio, and now the host and the guests seems to boot at without a problem. Regards, Andres From abo at root.snowtree.se Mon Nov 23 22:52:01 2009 From: abo at root.snowtree.se (Alexander =?ISO-8859-1?Q?Bostr=F6m?=) Date: Mon, 23 Nov 2009 23:52:01 +0100 Subject: [fedora-virt] Routing to guests In-Reply-To: <4B0AA214.9050407@pricom.com.au> References: <4B0AA214.9050407@pricom.com.au> Message-ID: <1259016721.3567.31.camel@tempo.alexander.bostrom.net> tis 2009-11-24 klockan 01:54 +1100 skrev Philip Rhoades: > I can ssh from/to the host/guest OK but how do I set up a route (or > whatever is necessary) so that another machine: > > eth0: 192.168.0.12 > > can ssh to the guest? I think the solution you're really looking for is to create a bridge that's attached to eth0 directly. The host would then perform no routing at all and your guest would get an IP address directly from your ADSL router. There's no convenient way to do it with libvirt and NetworkManager yet, though. /abo (That's my nick, I've had it for a long time! I didn't realise it was a bad word until i visited Sydney. *pout*) From rhdrop at remcc.org Tue Nov 24 01:03:20 2009 From: rhdrop at remcc.org (Robert Thiem) Date: Tue, 24 Nov 2009 11:03:20 +1000 (EST) Subject: [fedora-virt] Routing to guests In-Reply-To: <20091123155536.52B24619E13@hormel.redhat.com> References: <20091123155536.52B24619E13@hormel.redhat.com> Message-ID: <57736.203.25.1.208.1259024600.squirrel@gw.remcc.org> > From: Philip Rhoades > I can ssh from/to the host/guest OK but how do I set up a route (or > whatever is necessary) so that another machine: > eth0: 192.168.0.12 > can ssh to the guest? - "ssh 192.168.122.68" gives "no route to host" - > http://docs.fedoraproject.org/virtualization-guide/f12/en-US/html/ but > the problem does not seem to be covered there. Alexander is correct in saying that bridging would allow you to do that. There are two networking discussed in the guide. The first is a NAT (network address translation), in which the guests are given "private" ip addresses and any outbound traffic appears to be coming from the host machine's IP address. This is the same as the setup on your ADSL router where the internal network machines get addresses of 192.168.x.x but the internet sees your requests as coming from the IP address of your router. There should be lots of documentation in linux firewalling guides under sections on NAT (or possibly called IP Masquerading in some). Have a look at these for information on port forwarding to reveal services inside the virtual (such as ssh). The other option is bridging. This shares the physical network interface of the host with the guest. In this case the VM acts as though it's a machine plugged into the same subnet as the host, its services are accessible like those of the host and it's as vulnerable to attack as the host. Robert From markmc at redhat.com Tue Nov 24 07:37:35 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Tue, 24 Nov 2009 07:37:35 +0000 Subject: [fedora-virt] Routing to guests In-Reply-To: <57736.203.25.1.208.1259024600.squirrel@gw.remcc.org> References: <20091123155536.52B24619E13@hormel.redhat.com> <57736.203.25.1.208.1259024600.squirrel@gw.remcc.org> Message-ID: <1259048255.8935.12.camel@blaa> On Tue, 2009-11-24 at 11:03 +1000, Robert Thiem wrote: > The other option is bridging. This shares the physical network interface > of the host with the guest. In this case the VM acts as though it's a > machine plugged into the same subnet as the host, its services are > accessible like those of the host and it's as vulnerable to attack as the > host. ... and instructions on how to configure this mode are available here: http://wiki.libvirt.org/page/Networking#Fedora.2FRHEL_Bridging Cheers, Mark. From gianluca.cecchi at gmail.com Tue Nov 24 10:53:46 2009 From: gianluca.cecchi at gmail.com (Gianluca Cecchi) Date: Tue, 24 Nov 2009 11:53:46 +0100 Subject: [fedora-virt] F12 and ksm/ksmtuned: info required... Message-ID: <561c252c0911240253u7ed558dfk947faf581c311181@mail.gmail.com> Hello, F12 x86_64 with Qemu/KVM booted yesterday evening. It has two guests with CentOS 5.3 x86_64 configured to auto-start and now they are running... I see this morning, some minutes ago: [root at virtfed ~]# service ksm status ksm is not running My config: [root at virtfed ~]# chkconfig --list|grep ksm ksm 0:off 1:off 2:on 3:on 4:on 5:on 6:off ksmtuned 0:off 1:off 2:on 3:on 4:on 5:on 6:off [root at virtfed ~]# cat /sys/kernel/mm/ksm/pages_shared 3786 Then I start two other guests, both with CentOS 5.4 x86_64. After a while: [root at virtfed ~]# cat /sys/kernel/mm/ksm/pages_shared 4752 [root at virtfed ~]# cat /sys/kernel/mm/ksm/pages_shared 5219 ... [root at virtfed ~]# cat /sys/kernel/mm/ksm/pages_shared 90895 and now, without any action at my side: [root at virtfed ~]# service ksm status ksm is running Is this normal/expected? Any deeper doc/link about ksm/ksmtuned logic? Is it ksmtuned that spins ksm as it is needed? Any log file to check/configure? I read https://fedoraproject.org/wiki/Features/KSM but probably I'm missing something... Thanks, Gianluca From markmc at redhat.com Tue Nov 24 10:55:56 2009 From: markmc at redhat.com (Mark McLoughlin) Date: Tue, 24 Nov 2009 10:55:56 +0000 Subject: [fedora-virt] F12 and ksm/ksmtuned: info required... In-Reply-To: <561c252c0911240253u7ed558dfk947faf581c311181@mail.gmail.com> References: <561c252c0911240253u7ed558dfk947faf581c311181@mail.gmail.com> Message-ID: <1259060156.8935.24.camel@blaa> On Tue, 2009-11-24 at 11:53 +0100, Gianluca Cecchi wrote: > [root at virtfed ~]# service ksm status > ksm is not running 'bash -x /etc/init.d/ksm status' might help figure out why it's returning 'not running' Thanks, Mark. From berrange at redhat.com Tue Nov 24 11:13:23 2009 From: berrange at redhat.com (Daniel P. Berrange) Date: Tue, 24 Nov 2009 11:13:23 +0000 Subject: [fedora-virt] Working sound in KVM with SDL display on Fedora 12 ? Message-ID: <20091124111323.GD8427@redhat.com> I've been working on GTK-VNC to make it able to receive the audio stream from the remote VNC server, and decided to compare the quality against the current KVM host audio support. At which point I discovered that, AFAICT, the latter does not work in the slightest. I know that when running KVM guests with VNC under libvirt, you will not get any sound because I disabled that bit of code :-) It is guests using SDL graphics I'm interested in. eg, running a really simple guest like /usr/bin/qemu-kvm -m 700 -smp 1 -boot c -drive file=/home/berrange/f11i686.img,if=virtio,index=0,boot=on -vga cirrus -soundhw ac97 -sdl I have tested both ac97, and es1370 sound cards. I have tested with the default QEMU host audio backend (pulseaudio), and with also tested with each of the following settings on the host QEMU_AUDIO_DRV=pa QEMU_AUDIO_DRV=alsa QEMU_AUDIO_DRV=sdl QEMU_AUDIO_DRV=sdl SDL_AUDIODRIVER=pa QEMU_AUDIO_DRV=sdl SDL_AUDIODRIVER=alsa NB, they all ultimately go to pulseaudio on the host since it has its alsa proxy installed All of them appear to be more or less just as bad as each other. Inside the guest I'm running a boring old F11 i686 guest. I removed pulseaudio inside the guest and run 'mpg321' on the native alsa drivers. If you include the '-v' flag to mpg321, it shows its playback progress. What I'm seeing is that for the first 10 seconds or so its timer runs at 1/2 the actual speed, and then it speeds up exponentially taking a mere 10 seconds for the remaining 2 minutes of the mp3. Does anyone successfully have KVM on Fedora 12 playing audio when using the SDL display ? AFAICT, audio is just totally & utterly fubar, with SDL and any of the host audio drivers. So if anyone has it working reliably please tell me what, if any, QEMU_AUDIO_DRV / SDL_AUDIODRIVER env variables you have set, what guest OS you use, and what program you are using for playback in the guest. Also on the host, whether you have pulse audio (the default for F12), or are using native ALSA directly. On the plus side, by comparison, my code for streaming audio to GTK-VNC is managing to playback at pretty much native speed, with only a few dropouts in buffering 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 danken at redhat.com Tue Nov 24 11:50:18 2009 From: danken at redhat.com (Dan Kenigsberg) Date: Tue, 24 Nov 2009 13:50:18 +0200 Subject: [fedora-virt] F12 and ksm/ksmtuned: info required... In-Reply-To: <561c252c0911240253u7ed558dfk947faf581c311181@mail.gmail.com> References: <561c252c0911240253u7ed558dfk947faf581c311181@mail.gmail.com> Message-ID: <20091124115017.GA27114@redhat.com> On Tue, Nov 24, 2009 at 11:53:46AM +0100, Gianluca Cecchi wrote: > Hello, > F12 x86_64 with Qemu/KVM booted yesterday evening. > It has two guests with CentOS 5.3 x86_64 configured to auto-start and > now they are running... > I see this morning, some minutes ago: > [root at virtfed ~]# service ksm status > ksm is not running > > My config: > [root at virtfed ~]# chkconfig --list|grep ksm > ksm 0:off 1:off 2:on 3:on 4:on 5:on 6:off > ksmtuned 0:off 1:off 2:on 3:on 4:on 5:on 6:off > > [root at virtfed ~]# cat /sys/kernel/mm/ksm/pages_shared > 3786 > > Then I start two other guests, both with CentOS 5.4 x86_64. > After a while: > [root at virtfed ~]# cat /sys/kernel/mm/ksm/pages_shared > 4752 > [root at virtfed ~]# cat /sys/kernel/mm/ksm/pages_shared > 5219 > ... > [root at virtfed ~]# cat /sys/kernel/mm/ksm/pages_shared > 90895 > > and now, without any action at my side: > [root at virtfed ~]# service ksm status > ksm is running > > Is this normal/expected? > Any deeper doc/link about ksm/ksmtuned logic? > Is it ksmtuned that spins ksm as it is needed? Any log file to check/configure? > I read https://fedoraproject.org/wiki/Features/KSM but probably I'm > missing something... yes, the purpose of ksmtuned is to tune ksm parameters, or stop it altogether if it is not needed. In case your two virtual machines required less than available memeory, ksm is not used. If ksmtune senses that memory stress has risen, it fires up ksm again. I don't think this is written anywhere but in the code.. We probably should add something to the "User experience" section to that feature page. How about: Fedora's kvm comes with 2 services controlling the behavior of ksm. One, simply called ksm, is just a nice means to start and stop ksm's kernel thread. The other, called ksmtuned, controls the first service and tunes its parameters according to the memory stress that is generated by KVM virtual machines. > > Thanks, > Gianluca > > _______________________________________________ > Fedora-virt mailing list > Fedora-virt at redhat.com > https://www.redhat.com/mailman/listinfo/fedora-virt From gianluca.cecchi at gmail.com Tue Nov 24 12:01:02 2009 From: gianluca.cecchi at gmail.com (Gianluca Cecchi) Date: Tue, 24 Nov 2009 13:01:02 +0100 Subject: [fedora-virt] F12 and ksm/ksmtuned: info required... In-Reply-To: <20091124115017.GA27114@redhat.com> References: <561c252c0911240253u7ed558dfk947faf581c311181@mail.gmail.com> <20091124115017.GA27114@redhat.com> Message-ID: <561c252c0911240401n74f322c4vae1edd3f7d483d3e@mail.gmail.com> On Tue, Nov 24, 2009 at 12:50 PM, Dan Kenigsberg wrote: > > yes, the purpose of ksmtuned is to tune ksm parameters, or stop it > altogether if it is not needed. In case your two virtual machines > required less than available memeory, ksm is not used. If ksmtune senses > that memory stress has risen, it fires up ksm again. > > I don't think this is written anywhere but in the code.. We probably > should add something to the "User experience" section to that feature page. > How about: > > Fedora's kvm comes with 2 services controlling the behavior of ksm. One, > simply called ksm, is just a nice means to start and stop ksm's kernel > thread. The other, called ksmtuned, controls the first service and tunes > its parameters according to the memory stress that is generated by KVM > virtual machines. > It sounds good; I would only add something like this: "In case of need, e.g. in minor load situations, ksmtuned can also stop ksm service at all. Later, if ksmtuned senses that memory stress has risen, it will fire up ksm again." So that one knows that to monitor the status of ksm is not a good idea in general..... Any log configurable to see/trace ksmtuned decisions (start ksm, stop ksm, increase pages...)? Gianluca From kashyapc at fedoraproject.org Tue Nov 24 13:13:42 2009 From: kashyapc at fedoraproject.org (kashyap chamarthy) Date: Tue, 24 Nov 2009 18:43:42 +0530 Subject: [fedora-virt] F12 and ksm/ksmtuned: info required... In-Reply-To: <561c252c0911240401n74f322c4vae1edd3f7d483d3e@mail.gmail.com> References: <561c252c0911240253u7ed558dfk947faf581c311181@mail.gmail.com> <20091124115017.GA27114@redhat.com> <561c252c0911240401n74f322c4vae1edd3f7d483d3e@mail.gmail.com> Message-ID: <4B0BDC06.8000901@fedoraproject.org> On 11/24/2009 05:31 PM, Gianluca Cecchi wrote: > On Tue, Nov 24, 2009 at 12:50 PM, Dan Kenigsberg wrote: >> >> yes, the purpose of ksmtuned is to tune ksm parameters, or stop it >> altogether if it is not needed. In case your two virtual machines >> required less than available memeory, ksm is not used. If ksmtune senses >> that memory stress has risen, it fires up ksm again. >> >> I don't think this is written anywhere but in the code.. We probably >> should add something to the "User experience" section to that feature page. >> How about: >> >> Fedora's kvm comes with 2 services controlling the behavior of ksm. One, >> simply called ksm, is just a nice means to start and stop ksm's kernel >> thread. The other, called ksmtuned, controls the first service and tunes >> its parameters according to the memory stress that is generated by KVM >> virtual machines. >> > > It sounds good; I would only add something like this: > > "In case of need, e.g. in minor load situations, ksmtuned can also > stop ksm service at all. > Later, if ksmtuned senses that memory stress has risen, it will fire > up ksm again." > > So that one knows that to monitor the status of ksm is not a good idea > in general..... > Any log configurable to see/trace ksmtuned decisions (start ksm, stop > ksm, increase pages...)? yes, configuratble log/trace info for ksm/ksmtuned would be useful. I see this for configuration stuff - /etc/ksmtuned.conf /kashyap > > Gianluca > > _______________________________________________ > Fedora-virt mailing list > Fedora-virt at redhat.com > https://www.redhat.com/mailman/listinfo/fedora-virt > From danken at redhat.com Tue Nov 24 14:30:41 2009 From: danken at redhat.com (Dan Kenigsberg) Date: Tue, 24 Nov 2009 16:30:41 +0200 Subject: [fedora-virt] F12 and ksm/ksmtuned: info required... In-Reply-To: <4B0BDC06.8000901@fedoraproject.org> References: <561c252c0911240253u7ed558dfk947faf581c311181@mail.gmail.com> <20091124115017.GA27114@redhat.com> <561c252c0911240401n74f322c4vae1edd3f7d483d3e@mail.gmail.com> <4B0BDC06.8000901@fedoraproject.org> Message-ID: <20091124143041.GA28975@redhat.com> On Tue, Nov 24, 2009 at 06:43:42PM +0530, kashyap chamarthy wrote: > On 11/24/2009 05:31 PM, Gianluca Cecchi wrote: > >On Tue, Nov 24, 2009 at 12:50 PM, Dan Kenigsberg wrote: > >> > >>yes, the purpose of ksmtuned is to tune ksm parameters, or stop it > >>altogether if it is not needed. In case your two virtual machines > >>required less than available memeory, ksm is not used. If ksmtune senses > >>that memory stress has risen, it fires up ksm again. > >> > >>I don't think this is written anywhere but in the code.. We probably > >>should add something to the "User experience" section to that feature page. > >>How about: > >> > >>Fedora's kvm comes with 2 services controlling the behavior of ksm. One, > >>simply called ksm, is just a nice means to start and stop ksm's kernel > >>thread. The other, called ksmtuned, controls the first service and tunes > >>its parameters according to the memory stress that is generated by KVM > >>virtual machines. > >> > > > >It sounds good; I would only add something like this: > > > >"In case of need, e.g. in minor load situations, ksmtuned can also > >stop ksm service at all. > >Later, if ksmtuned senses that memory stress has risen, it will fire > >up ksm again." > > > >So that one knows that to monitor the status of ksm is not a good idea > >in general..... > >Any log configurable to see/trace ksmtuned decisions (start ksm, stop > >ksm, increase pages...)? > > yes, configuratble log/trace info for ksm/ksmtuned would be useful. > > I see this for configuration stuff - /etc/ksmtuned.conf > > /kashyap You'd have to apply the following patch to /usr/sbin/ksmtuned for it to easily log into /var/log/ksmtuned. You'd have to add two lines to /etc/ksmtuned.conf for this to actually work. Regards, Dan. From danken at redhat.com Tue Nov 24 14:31:34 2009 From: danken at redhat.com (Dan Kenigsberg) Date: Tue, 24 Nov 2009 16:31:34 +0200 Subject: [fedora-virt] [PATCH] optionally send debug to /var/log/ksmtuned In-Reply-To: <20091124143041.GA28975@redhat.com> References: <20091124143041.GA28975@redhat.com> Message-ID: <1259073094-32074-1-git-send-email-danken@redhat.com> --- ksmtuned | 25 ++++++++++++++++--------- ksmtuned.conf | 5 +++++ 2 files changed, 21 insertions(+), 9 deletions(-) diff --git a/ksmtuned b/ksmtuned index 7da8b68..decf41d 100644 --- a/ksmtuned +++ b/ksmtuned @@ -21,6 +21,13 @@ if [ -f /etc/ksmtuned.conf ]; then . /etc/ksmtuned.conf fi +debug() { + if [ -n "$DEBUG" ]; then + s="`/bin/date`: $*" + [ -n "$LOGFILE" ] && echo "$s" >> "$LOGFILE" || echo "$s" + fi +} + KSM_MONITOR_INTERVAL=${KSM_MONITOR_INTERVAL:-60} KSM_NPAGES_BOOST=${KSM_NPAGES_BOOST:-300} KSM_NPAGES_DECAY=${KSM_NPAGES_DECAY:--50} @@ -35,17 +42,17 @@ KSM_THRES_COEF=${KSM_THRES_COEF:-20} KSM_THRES_CONST=${KSM_THRES_CONST:-2048} total=`awk '/^MemTotal:/ {print $2}' /proc/meminfo` -[ -n "$DEBUG" ] && echo total $total +debug total $total npages=0 sleep=$[KSM_SLEEP_MSEC * 16 * 1024 * 1024 / total] [ $sleep -le 10 ] && sleep=10 -[ -n "$DEBUG" ] && echo sleep $sleep +debug sleep $sleep thres=$[total * KSM_THRES_COEF / 100] if [ $KSM_THRES_CONST -gt $thres ]; then thres=$KSM_THRES_CONST fi -[ -n "$DEBUG" ] && echo thres $thres +debug thres $thres KSMCTL () { case x$1 in @@ -89,22 +96,22 @@ adjust () { local free committed free=`free_memory` committed=`committed_memory` - [ -n "$DEBUG" ] && echo committed $committed free $free + debug committed $committed free $free if [ $[committed + thres] -lt $total -a $free -gt $thres ]; then KSMCTL stop - [ -n "$DEBUG" ] && echo "$[committed + thres] < $total and free > $thres, stop ksm" + debug "$[committed + thres] < $total and free > $thres, stop ksm" return 1 fi - [ -n "$DEBUG" ] && echo "$[committed + thres] > $total, start ksm" + debug "$[committed + thres] > $total, start ksm" if [ $free -lt $thres ]; then npages=`increase_npages $KSM_NPAGES_BOOST` - [ -n "$DEBUG" ] && echo "$free < $thres, boost" + debug "$free < $thres, boost" else npages=`increase_npages $KSM_NPAGES_DECAY` - [ -n "$DEBUG" ] && echo "$free > $thres, decay" + debug "$free > $thres, decay" fi KSMCTL start $npages $sleep - [ -n "$DEBUG" ] && echo "KSMCTL start $npages $sleep" + debug "KSMCTL start $npages $sleep" return 0 } diff --git a/ksmtuned.conf b/ksmtuned.conf index 87b9178..fc4518c 100644 --- a/ksmtuned.conf +++ b/ksmtuned.conf @@ -14,3 +14,8 @@ # KSM_THRES_COEF=20 # KSM_THRES_CONST=2048 + +# uncomment the following if you want ksmtuned debug info + +# LOGFILE=/var/log/ksmtuned +# DEBUG=1 -- 1.6.5.2 From kashyapc at fedoraproject.org Tue Nov 24 15:43:40 2009 From: kashyapc at fedoraproject.org (Kashyap Chamarthy) Date: Tue, 24 Nov 2009 21:13:40 +0530 Subject: [fedora-virt] F12 and ksm/ksmtuned: info required... In-Reply-To: <20091124143041.GA28975@redhat.com> References: <561c252c0911240253u7ed558dfk947faf581c311181@mail.gmail.com> <20091124115017.GA27114@redhat.com> <561c252c0911240401n74f322c4vae1edd3f7d483d3e@mail.gmail.com> <4B0BDC06.8000901@fedoraproject.org> <20091124143041.GA28975@redhat.com> Message-ID: <4B0BFF2C.2070407@fedoraproject.org> On 11/24/2009 08:00 PM, Dan Kenigsberg wrote: > On Tue, Nov 24, 2009 at 06:43:42PM +0530, kashyap chamarthy wrote: >> On 11/24/2009 05:31 PM, Gianluca Cecchi wrote: >>> On Tue, Nov 24, 2009 at 12:50 PM, Dan Kenigsberg wrote: >>>> >>>> yes, the purpose of ksmtuned is to tune ksm parameters, or stop it >>>> altogether if it is not needed. In case your two virtual machines >>>> required less than available memeory, ksm is not used. If ksmtune senses >>>> that memory stress has risen, it fires up ksm again. >>>> >>>> I don't think this is written anywhere but in the code.. We probably >>>> should add something to the "User experience" section to that feature page. >>>> How about: >>>> >>>> Fedora's kvm comes with 2 services controlling the behavior of ksm. One, >>>> simply called ksm, is just a nice means to start and stop ksm's kernel >>>> thread. The other, called ksmtuned, controls the first service and tunes >>>> its parameters according to the memory stress that is generated by KVM >>>> virtual machines. >>>> >>> >>> It sounds good; I would only add something like this: >>> >>> "In case of need, e.g. in minor load situations, ksmtuned can also >>> stop ksm service at all. >>> Later, if ksmtuned senses that memory stress has risen, it will fire >>> up ksm again." >>> >>> So that one knows that to monitor the status of ksm is not a good idea >>> in general..... >>> Any log configurable to see/trace ksmtuned decisions (start ksm, stop >>> ksm, increase pages...)? >> >> yes, configuratble log/trace info for ksm/ksmtuned would be useful. >> >> I see this for configuration stuff - /etc/ksmtuned.conf >> >> /kashyap > > You'd have to apply the following patch to /usr/sbin/ksmtuned for it to > easily log into /var/log/ksmtuned. You'd have to add two lines to > /etc/ksmtuned.conf for this to actually work. thanks a lot. will try it. /kashyap > > Regards, > > Dan. > From kashyapc at fedoraproject.org Tue Nov 24 18:14:01 2009 From: kashyapc at fedoraproject.org (Kashyap Chamarthy) Date: Tue, 24 Nov 2009 23:44:01 +0530 Subject: [fedora-virt] F12 and ksm/ksmtuned: info required... In-Reply-To: <4B0BFF2C.2070407@fedoraproject.org> References: <561c252c0911240253u7ed558dfk947faf581c311181@mail.gmail.com> <20091124115017.GA27114@redhat.com> <561c252c0911240401n74f322c4vae1edd3f7d483d3e@mail.gmail.com> <4B0BDC06.8000901@fedoraproject.org> <20091124143041.GA28975@redhat.com> <4B0BFF2C.2070407@fedoraproject.org> Message-ID: <4B0C2269.7020307@fedoraproject.org> On 11/24/2009 09:13 PM, Kashyap Chamarthy wrote: > On 11/24/2009 08:00 PM, Dan Kenigsberg wrote: >> On Tue, Nov 24, 2009 at 06:43:42PM +0530, kashyap chamarthy wrote: >>> On 11/24/2009 05:31 PM, Gianluca Cecchi wrote: >>>> On Tue, Nov 24, 2009 at 12:50 PM, Dan Kenigsberg >>>> wrote: >>>>> >>>>> yes, the purpose of ksmtuned is to tune ksm parameters, or stop it >>>>> altogether if it is not needed. In case your two virtual machines >>>>> required less than available memeory, ksm is not used. If ksmtune >>>>> senses >>>>> that memory stress has risen, it fires up ksm again. >>>>> >>>>> I don't think this is written anywhere but in the code.. We probably >>>>> should add something to the "User experience" section to that >>>>> feature page. >>>>> How about: >>>>> >>>>> Fedora's kvm comes with 2 services controlling the behavior of ksm. >>>>> One, >>>>> simply called ksm, is just a nice means to start and stop ksm's kernel >>>>> thread. The other, called ksmtuned, controls the first service and >>>>> tunes >>>>> its parameters according to the memory stress that is generated by KVM >>>>> virtual machines. >>>>> >>>> >>>> It sounds good; I would only add something like this: >>>> >>>> "In case of need, e.g. in minor load situations, ksmtuned can also >>>> stop ksm service at all. >>>> Later, if ksmtuned senses that memory stress has risen, it will fire >>>> up ksm again." >>>> >>>> So that one knows that to monitor the status of ksm is not a good idea >>>> in general..... >>>> Any log configurable to see/trace ksmtuned decisions (start ksm, stop >>>> ksm, increase pages...)? >>> >>> yes, configuratble log/trace info for ksm/ksmtuned would be useful. >>> >>> I see this for configuration stuff - /etc/ksmtuned.conf >>> >>> /kashyap >> >> You'd have to apply the following patch to /usr/sbin/ksmtuned for it to >> easily log into /var/log/ksmtuned. You'd have to add two lines to >> /etc/ksmtuned.conf for this to actually work. > > thanks a lot. will try it. nice..after applying the patch and uncommenting the below lines in /etc/ksmtuned.conf throws ksmtuned activity under/var/log/ksmtuned # LOGFILE=/var/log/ksmtuned # DEBUG=1 ------------------------------------------------- [root at foobox ~]# /etc/init.d/ksmtuned restart Stopping ksmtuned: [ OK ] Starting ksmtuned: [ OK ] [root at foobox ~]# cat /var/log/ksmtuned Tue Nov 24 22:40:59 IST 2009: total 4019072 Tue Nov 24 22:40:59 IST 2009: sleep 41 Tue Nov 24 22:40:59 IST 2009: thres 803814 Tue Nov 24 22:41:50 IST 2009: total 4019072 Tue Nov 24 22:41:50 IST 2009: sleep 41 Tue Nov 24 22:41:50 IST 2009: thres 803814 --------------------------------------------------- > > /kashyap > >> >> Regards, >> >> Dan. >> > > _______________________________________________ > Fedora-virt mailing list > Fedora-virt at redhat.com > https://www.redhat.com/mailman/listinfo/fedora-virt > From eb30750 at gmail.com Tue Nov 24 18:46:27 2009 From: eb30750 at gmail.com (Paul Lambert) Date: Tue, 24 Nov 2009 13:46:27 -0500 Subject: [fedora-virt] wireless network connectivity Message-ID: I need some feedback as to what could be the source of my wireless connectivity for FE-12. Previously, there were some serious problems with the wireless network software as well as the user interface using FE-12. In general, these issues have been corrected. The problem I now see is dropped connections. This could be due to failure to re-authenticate or something in the driver. I mostly see this problem when connecting to a Linksys WRT160N. However, once the problem is observed the FE-12 will not reconnect to any wireless router unless I logoff and then log back on. Or, simply reboot the computer. It should be noted that the WRT160N has been plagued by software problems. In fact, you can buy an entire new OS for these and many problems go away. I using v3 the latest from Linksys for this device. For the most part when I connect to a WRT54G I do not see these problems. Scenario: After bootup the computer will connect to the WRT160N and communicate as expected. The longer I stayed connected the slower the response gets. Eventually, when executing an "arp -a" command the system will return the MAC address with a instead of the IP address for the router. The computer's IP address 192.168.1.100 and the router is 10.10.10.1. At this point attempting to connect to a different router (WRT54G) does not work either. Only a logout or reboot corrects the problem. I am using WPA & WPA2 Personal authentication whereas the WRT54G uses WEP. I am hesitant to file a bug report due to the possibility the root problem is not in FE-12. However, the fact that the FE-12 network drive does not recover indicates something is not quite right. Thanks PJAL -------------- next part -------------- An HTML attachment was scrubbed... URL: From jmforbes at linuxtx.org Tue Nov 24 18:52:51 2009 From: jmforbes at linuxtx.org (Justin M. Forbes) Date: Tue, 24 Nov 2009 12:52:51 -0600 Subject: [fedora-virt] [PATCH] optionally send debug to /var/log/ksmtuned In-Reply-To: <1259073094-32074-1-git-send-email-danken@redhat.com> References: <20091124143041.GA28975@redhat.com> <1259073094-32074-1-git-send-email-danken@redhat.com> Message-ID: <20091124185251.GC24148@linuxtx.org> On Tue, Nov 24, 2009 at 04:31:34PM +0200, Dan Kenigsberg wrote: > --- > ksmtuned | 25 ++++++++++++++++--------- > ksmtuned.conf | 5 +++++ > 2 files changed, 21 insertions(+), 9 deletions(-) > > diff --git a/ksmtuned b/ksmtuned > index 7da8b68..decf41d 100644 > --- a/ksmtuned > +++ b/ksmtuned > @@ -21,6 +21,13 @@ if [ -f /etc/ksmtuned.conf ]; then > . /etc/ksmtuned.conf > fi > > +debug() { > + if [ -n "$DEBUG" ]; then > + s="`/bin/date`: $*" > + [ -n "$LOGFILE" ] && echo "$s" >> "$LOGFILE" || echo "$s" > + fi > +} > + > KSM_MONITOR_INTERVAL=${KSM_MONITOR_INTERVAL:-60} > KSM_NPAGES_BOOST=${KSM_NPAGES_BOOST:-300} > KSM_NPAGES_DECAY=${KSM_NPAGES_DECAY:--50} > @@ -35,17 +42,17 @@ KSM_THRES_COEF=${KSM_THRES_COEF:-20} > KSM_THRES_CONST=${KSM_THRES_CONST:-2048} > > total=`awk '/^MemTotal:/ {print $2}' /proc/meminfo` > -[ -n "$DEBUG" ] && echo total $total > +debug total $total > > npages=0 > sleep=$[KSM_SLEEP_MSEC * 16 * 1024 * 1024 / total] > [ $sleep -le 10 ] && sleep=10 > -[ -n "$DEBUG" ] && echo sleep $sleep > +debug sleep $sleep > thres=$[total * KSM_THRES_COEF / 100] > if [ $KSM_THRES_CONST -gt $thres ]; then > thres=$KSM_THRES_CONST > fi > -[ -n "$DEBUG" ] && echo thres $thres > +debug thres $thres > > KSMCTL () { > case x$1 in > @@ -89,22 +96,22 @@ adjust () { > local free committed > free=`free_memory` > committed=`committed_memory` > - [ -n "$DEBUG" ] && echo committed $committed free $free > + debug committed $committed free $free > if [ $[committed + thres] -lt $total -a $free -gt $thres ]; then > KSMCTL stop > - [ -n "$DEBUG" ] && echo "$[committed + thres] < $total and free > $thres, stop ksm" > + debug "$[committed + thres] < $total and free > $thres, stop ksm" > return 1 > fi > - [ -n "$DEBUG" ] && echo "$[committed + thres] > $total, start ksm" > + debug "$[committed + thres] > $total, start ksm" > if [ $free -lt $thres ]; then > npages=`increase_npages $KSM_NPAGES_BOOST` > - [ -n "$DEBUG" ] && echo "$free < $thres, boost" > + debug "$free < $thres, boost" > else > npages=`increase_npages $KSM_NPAGES_DECAY` > - [ -n "$DEBUG" ] && echo "$free > $thres, decay" > + debug "$free > $thres, decay" > fi > KSMCTL start $npages $sleep > - [ -n "$DEBUG" ] && echo "KSMCTL start $npages $sleep" > + debug "KSMCTL start $npages $sleep" > return 0 > } > > diff --git a/ksmtuned.conf b/ksmtuned.conf > index 87b9178..fc4518c 100644 > --- a/ksmtuned.conf > +++ b/ksmtuned.conf > @@ -14,3 +14,8 @@ > > # KSM_THRES_COEF=20 > # KSM_THRES_CONST=2048 > + > +# uncomment the following if you want ksmtuned debug info > + > +# LOGFILE=/var/log/ksmtuned > +# DEBUG=1 > -- Patch looks good, I will get it applied. I will also add a logrotate file so that this log doesn't grow unrestrained. Justin From erenoglu at gmail.com Tue Nov 24 23:03:20 2009 From: erenoglu at gmail.com (Emre Erenoglu) Date: Wed, 25 Nov 2009 00:03:20 +0100 Subject: [fedora-virt] Working sound in KVM with SDL display on Fedora 12 ? In-Reply-To: <20091124111323.GD8427@redhat.com> References: <20091124111323.GD8427@redhat.com> Message-ID: On Tue, Nov 24, 2009 at 12:13 PM, Daniel P. Berrange wrote: > > On the plus side, by comparison, my code for streaming audio to GTK-VNC is > managing to playback at pretty much native speed, with only a few dropouts > in buffering > > Hi Daniel, It won't answer your question, but just as a comment: I never got these things to work nicely and automatically. The only way I managed to get more or less reliable sound from my linux boxes were to have Pulseaudio inside the guest and use the host pulseaudio daemon over network. Emre -------------- next part -------------- An HTML attachment was scrubbed... URL: From crobinso at redhat.com Wed Nov 25 14:50:40 2009 From: crobinso at redhat.com (Cole Robinson) Date: Wed, 25 Nov 2009 09:50:40 -0500 Subject: [fedora-virt] wireless network connectivity In-Reply-To: References: Message-ID: <4B0D4440.4010107@redhat.com> On 11/24/2009 01:46 PM, Paul Lambert wrote: > I need some feedback as to what could be the source of my wireless > connectivity for FE-12. Previously, there were some serious problems with > the wireless network software as well as the user interface using FE-12. In > general, these issues have been corrected. > > The problem I now see is dropped connections. This could be due to failure > to re-authenticate or something in the driver. I mostly see this problem > when connecting to a Linksys WRT160N. However, once the problem is observed > the FE-12 will not reconnect to any wireless router unless I logoff and then > log back on. Or, simply reboot the computer. > > It should be noted that the WRT160N has been plagued by software problems. > In fact, you can buy an entire new OS for these and many problems go away. > I using v3 the latest from Linksys for this device. For the most part when > I connect to a WRT54G I do not see these problems. > > Scenario: > After bootup the computer will connect to the WRT160N and communicate as > expected. The longer I stayed connected the slower the response gets. > Eventually, when executing an "arp -a" command the system will return the > MAC address with a instead of the IP address for the router. > The computer's IP address 192.168.1.100 and the router is 10.10.10.1. At > this point attempting to connect to a different router (WRT54G) does not > work either. Only a logout or reboot corrects the problem. > > I am using WPA & WPA2 Personal authentication whereas the WRT54G uses WEP. > I am hesitant to file a bug report due to the possibility the root problem > is not in FE-12. However, the fact that the FE-12 network drive does not > recover indicates something is not quite right. > Wrong list? This doesn't seem virt related, maybe try fedora-list instead. - Cole From gianluca.cecchi at gmail.com Mon Nov 30 14:00:39 2009 From: gianluca.cecchi at gmail.com (Gianluca Cecchi) Date: Mon, 30 Nov 2009 15:00:39 +0100 Subject: [fedora-virt] some vm die in F12 Message-ID: <561c252c0911300600r3edd52a3h175c81ff25e30911@mail.gmail.com> Hello, I had an F11 host with virt-preview repo configured and I have used it a lot without particular problems. Now I'm on F12 on the same host after upgrade and I'm experiencing guests' death caused by oom killer. Example. - all VM (4) are qemu/kvm and x86_64 - boot of host on 26/11 about 19:20. - two centos 5.3 guests are configured to startup automatically and they indeed start In messages I see this on their startup: Nov 26 19:23:01 virtfed kernel: kvm: 3458: cpu0 unimplemented perfctr wrmsr: 0x186 data 0x130079 Nov 26 19:23:01 virtfed kernel: kvm: 3458: cpu0 unimplemented perfctr wrmsr: 0xc1 data 0xffcfadc8 Nov 26 19:23:01 virtfed kernel: kvm: 3458: cpu0 unimplemented perfctr wrmsr: 0x186 data 0x530079 Nov 26 19:23:01 virtfed kernel: kvm: 3458: cpu1 unimplemented perfctr wrmsr: 0x186 data 0x130079 Nov 26 19:23:01 virtfed kernel: kvm: 3458: cpu1 unimplemented perfctr wrmsr: 0xc1 data 0xffcfadc8 Nov 26 19:23:01 virtfed kernel: kvm: 3458: cpu1 unimplemented perfctr wrmsr: 0x186 data 0x530079 Nov 26 19:23:01 virtfed kernel: kvm: 3501: cpu0 unimplemented perfctr wrmsr: 0x186 data 0x130079 Nov 26 19:23:01 virtfed kernel: kvm: 3501: cpu0 unimplemented perfctr wrmsr: 0xc1 data 0xffcfadc8 Nov 26 19:23:01 virtfed kernel: kvm: 3501: cpu0 unimplemented perfctr wrmsr: 0x186 data 0x530079 Nov 26 19:23:01 virtfed kernel: kvm: 3501: cpu1 unimplemented perfctr wrmsr: 0x186 data 0x130079 Nov 26 19:23:36 virtfed kernel: kvm: emulating exchange as write - On 27/11 at 12:03 I start other two guests in centos 5.4 Again I get rows similar to the above ones plus: Nov 27 12:03:19 virtfed kernel: __ratelimit: 2 callbacks suppressed The day after at 06:55 Nov 28 03:27:06 virtfed logrotate: ALERT exited abnormally with [1] Nov 28 06:55:01 virtfed kernel: qemu-kvm invoked oom-killer: gfp_mask=0x201da, order=0, oomkilladj=0 Nov 28 06:55:01 virtfed kernel: qemu-kvm cpuset=/ mems_allowed=0 Nov 28 06:55:01 virtfed kernel: Pid: 405, comm: qemu-kvm Tainted: G W 2.6.31.5-127.fc12.x86_64 #1 Nov 28 06:55:01 virtfed kernel: Call Trace: Nov 28 06:55:01 virtfed kernel: [] ? cpuset_print_task_mems_allowed+0x91/0x9d Nov 28 06:55:01 virtfed kernel: [] oom_kill_process+0x98/0x256 Nov 28 06:55:01 virtfed kernel: [] ? select_bad_process+0xa3/0x102 Nov 28 06:55:01 virtfed kernel: [] __out_of_memory+0x8a/0x99 Nov 28 06:55:01 virtfed kernel: [] out_of_memory+0x163/0x195 Nov 28 06:55:01 virtfed kernel: [] __alloc_pages_nodemask+0x491/0x584 Nov 28 06:55:01 virtfed kernel: [] alloc_pages_current+0x95/0x9e Nov 28 06:55:01 virtfed kernel: [] __page_cache_alloc+0x5f/0x61 Nov 28 06:55:01 virtfed kernel: [] __do_page_cache_readahead+0x98/0x176 Nov 28 06:55:01 virtfed kernel: [] ra_submit+0x21/0x25 Nov 28 06:55:01 virtfed kernel: [] filemap_fault+0x193/0x317 Nov 28 06:55:01 virtfed kernel: [] __do_fault+0x54/0x3c4 Nov 28 06:55:01 virtfed kernel: [] ? read_tsc+0x9/0x1b Nov 28 06:55:01 virtfed kernel: [] handle_mm_fault+0x2f6/0x705 Nov 28 06:55:01 virtfed kernel: [] ? __hrtimer_start_range_ns+0x2bb/0x2cd Nov 28 06:55:01 virtfed kernel: [] do_page_fault+0x281/0x299 Nov 28 06:55:01 virtfed kernel: [] page_fault+0x25/0x30 Nov 28 06:55:01 virtfed kernel: Mem-Info: Nov 28 06:55:01 virtfed kernel: Node 0 DMA per-cpu: Nov 28 06:55:01 virtfed kernel: CPU 0: hi: 0, btch: 1 usd: 0 Nov 28 06:55:01 virtfed kernel: CPU 1: hi: 0, btch: 1 usd: 0 Nov 28 06:55:01 virtfed kernel: CPU 2: hi: 0, btch: 1 usd: 0 Nov 28 06:55:01 virtfed kernel: CPU 3: hi: 0, btch: 1 usd: 0 Nov 28 06:55:01 virtfed kernel: CPU 4: hi: 0, btch: 1 usd: 0 Nov 28 06:55:01 virtfed kernel: CPU 5: hi: 0, btch: 1 usd: 0 Nov 28 06:55:01 virtfed kernel: CPU 6: hi: 0, btch: 1 usd: 0 Nov 28 06:55:01 virtfed kernel: CPU 7: hi: 0, btch: 1 usd: 0 Nov 28 06:55:01 virtfed kernel: Node 0 DMA32 per-cpu: Nov 28 06:55:01 virtfed kernel: CPU 0: hi: 186, btch: 31 usd: 86 Nov 28 06:55:01 virtfed kernel: CPU 1: hi: 186, btch: 31 usd: 136 Nov 28 06:55:01 virtfed kernel: CPU 2: hi: 186, btch: 31 usd: 51 Nov 28 06:55:01 virtfed kernel: CPU 3: hi: 186, btch: 31 usd: 126 Nov 28 06:55:01 virtfed kernel: CPU 4: hi: 186, btch: 31 usd: 86 Nov 28 06:55:01 virtfed kernel: CPU 5: hi: 186, btch: 31 usd: 34 Nov 28 06:55:01 virtfed kernel: CPU 6: hi: 186, btch: 31 usd: 171 Nov 28 06:55:01 virtfed kernel: CPU 7: hi: 186, btch: 31 usd: 169 Nov 28 06:55:01 virtfed kernel: Node 0 Normal per-cpu: Nov 28 06:55:01 virtfed kernel: CPU 0: hi: 186, btch: 31 usd: 173 Nov 28 06:55:01 virtfed kernel: CPU 1: hi: 186, btch: 31 usd: 42 Nov 28 06:55:01 virtfed kernel: CPU 2: hi: 186, btch: 31 usd: 160 Nov 28 06:55:01 virtfed kernel: CPU 3: hi: 186, btch: 31 usd: 109 Nov 28 06:55:01 virtfed kernel: CPU 4: hi: 186, btch: 31 usd: 112 Nov 28 06:55:01 virtfed kernel: CPU 5: hi: 186, btch: 31 usd: 112 Nov 28 06:55:01 virtfed kernel: CPU 6: hi: 186, btch: 31 usd: 174 Nov 28 06:55:01 virtfed kernel: CPU 7: hi: 186, btch: 31 usd: 118 Nov 28 06:55:01 virtfed kernel: Active_anon:2537459 active_file:3 inactive_anon:364350 Nov 28 06:55:01 virtfed kernel: inactive_file:160 unevictable:12562 dirty:0 writeback:0 unstable:0 Nov 28 06:55:01 virtfed kernel: free:16256 slab:29833 mapped:7872 pagetables:10697 bounce:0 Nov 28 06:55:01 virtfed kernel: Node 0 DMA free:15824kB min:16kB low:20kB high:24kB active_anon:0kB inactive_anon:0kB active_file:0kB inactive_file:0kB unevictable:0kB present:15320kB pages_scanned:0 all_unreclaimable? yes Nov 28 06:55:01 virtfed kernel: lowmem_reserve[]: 0 3254 12092 12092 Nov 28 06:55:01 virtfed kernel: Node 0 DMA32 free:39060kB min:3784kB low:4728kB high:5676kB active_anon:2510452kB inactive_anon:502304kB active_file:12kB inactive_file:244kB unevictable:520kB present:3332660kB pages_scanned:120 all_unreclaimable? no Nov 28 06:55:01 virtfed kernel: lowmem_reserve[]: 0 0 8837 8837 Nov 28 06:55:01 virtfed kernel: Node 0 Normal free:10140kB min:10280kB low:12848kB high:15420kB active_anon:7639384kB inactive_anon:955096kB active_file:0kB inactive_file:396kB unevictable:49728kB present:9049596kB pages_scanned:0 all_unreclaimable? no Nov 28 06:55:01 virtfed kernel: lowmem_reserve[]: 0 0 0 0 Nov 28 06:55:01 virtfed kernel: Node 0 DMA: 2*4kB 1*8kB 2*16kB 1*32kB 2*64kB 0*128kB 1*256kB 0*512kB 1*1024kB 1*2048kB 3*4096kB = 15824kB Nov 28 06:55:01 virtfed kernel: Node 0 DMA32: 3793*4kB 2168*8kB 68*16kB 11*32kB 4*64kB 6*128kB 3*256kB 2*512kB 2*1024kB 0*2048kB 0*4096kB = 38820kB Nov 28 06:55:01 virtfed kernel: Node 0 Normal: 1948*4kB 5*8kB 0*16kB 0*32kB 2*64kB 0*128kB 1*256kB 0*512kB 0*1024kB 1*2048kB 0*4096kB = 10264kB Nov 28 06:55:01 virtfed kernel: 26703 total pagecache pages Nov 28 06:55:01 virtfed kernel: 18747 pages in swap cache Nov 28 06:55:01 virtfed kernel: Swap cache stats: add 1236041, delete 1217294, find 52896/69574 Nov 28 06:55:01 virtfed kernel: Free swap = 0kB Nov 28 06:55:01 virtfed kernel: Total swap = 4194296kB Nov 28 06:55:01 virtfed kernel: 3145727 pages RAM Nov 28 06:55:01 virtfed kernel: 63020 pages reserved Nov 28 06:55:01 virtfed kernel: 76044 pages shared Nov 28 06:55:01 virtfed kernel: 3053641 pages non-shared Nov 28 06:55:01 virtfed kernel: virbr0: port 3(vnet4) entering disabled state Nov 28 06:55:01 virtfed kernel: device vnet4 left promiscuous mode Nov 28 06:55:01 virtfed kernel: virbr0: port 3(vnet4) entering disabled state About 1 hour and a half after, at 08:14 I get the same for the other 5.4 vm. The same happens for other processes Any hints on this and on how to debug? Any other one? I don't know if it can be related, but I also tried to reboot disabling ksm and ksmtuned: [root at virtfed ~]# chkconfig --list ksmtuned ksmtuned 0:off 1:off 2:off 3:off 4:off 5:off 6:off [root at virtfed ~]# chkconfig --list ksm ksm 0:off 1:off 2:off 3:off 4:off 5:off 6:off getting the same problem... strangely, I get anyway this: [root at virtfed ~]# service ksm status ksm is running Why ksm is running? How can I configure a VM to be or not to be managed with ksm? Or is it an overall feature to enable/disable? From avi at redhat.com Mon Nov 30 14:38:11 2009 From: avi at redhat.com (Avi Kivity) Date: Mon, 30 Nov 2009 16:38:11 +0200 Subject: [fedora-virt] some vm die in F12 In-Reply-To: <561c252c0911300600r3edd52a3h175c81ff25e30911@mail.gmail.com> References: <561c252c0911300600r3edd52a3h175c81ff25e30911@mail.gmail.com> Message-ID: <4B13D8D3.4020805@redhat.com> On 11/30/2009 04:00 PM, Gianluca Cecchi wrote: > Hello, > I had an F11 host with virt-preview repo configured and I have used it > a lot without particular problems. > Now I'm on F12 on the same host after upgrade and I'm experiencing > guests' death caused by oom killer. > > Example. > - all VM (4) are qemu/kvm and x86_64 > - boot of host on 26/11 about 19:20. > - two centos 5.3 guests are configured to startup automatically and > they indeed start > How much memory have you allocated to guests, and how much memory do you have in your system? How much swap space? -- error compiling committee.c: too many arguments to function From gianluca.cecchi at gmail.com Mon Nov 30 14:48:24 2009 From: gianluca.cecchi at gmail.com (Gianluca Cecchi) Date: Mon, 30 Nov 2009 15:48:24 +0100 Subject: [fedora-virt] some vm die in F12 In-Reply-To: <4B13D8D3.4020805@redhat.com> References: <561c252c0911300600r3edd52a3h175c81ff25e30911@mail.gmail.com> <4B13D8D3.4020805@redhat.com> Message-ID: <561c252c0911300648i5091d17fn1e29eb541f7ecc64@mail.gmail.com> On Mon, Nov 30, 2009 at 3:38 PM, Avi Kivity wrote: > How much memory have you allocated to guests, and how much memory do you > have in your system? ?How much swap space? > > > The system has 10Gb of ram. Swap is configured for 4Gb 2 guests have 1Gb of ram configured 2 guests have 3,5Gb of ram configured. no other guests running Now I have the 2 guests with 1Gb of ram up and running and the situation is [root at virtfed log]# free total used free shared buffers cached Mem: 12330828 6005764 6325064 0 1099052 234008 -/+ buffers/cache: 4672704 7658124 Swap: 4194296 15880 4178416 In the mean time I have just installed sysstat to check memory usage during time... From gianluca.cecchi at gmail.com Mon Nov 30 14:58:45 2009 From: gianluca.cecchi at gmail.com (Gianluca Cecchi) Date: Mon, 30 Nov 2009 15:58:45 +0100 Subject: [fedora-virt] some vm die in F12 In-Reply-To: <561c252c0911300648i5091d17fn1e29eb541f7ecc64@mail.gmail.com> References: <561c252c0911300600r3edd52a3h175c81ff25e30911@mail.gmail.com> <4B13D8D3.4020805@redhat.com> <561c252c0911300648i5091d17fn1e29eb541f7ecc64@mail.gmail.com> Message-ID: <561c252c0911300658k2059e39aoa0e353f2aece377c@mail.gmail.com> On Mon, Nov 30, 2009 at 3:48 PM, Gianluca Cecchi wrote: > Now I have the 2 guests with 1Gb of ram up and running and the situation is > > [root at virtfed log]# free > ? ? ? ? ? ? total ? ? ? used ? ? ? free ? ? shared ? ?buffers ? ? cached > Mem: ? ? ?12330828 ? ?6005764 ? ?6325064 ? ? ? ? ?0 ? ?1099052 ? ? 234008 > -/+ buffers/cache: ? ?4672704 ? ?7658124 > Swap: ? ? ?4194296 ? ? ?15880 ? ?4178416 > > In the mean time I have just installed sysstat to check memory usage > during time... > I wrote about host having 10Gb of ram. Actually it has 12Gb of ram (as seen btw from the output of the "free" command). And after starting the other two guests (3,5Gb of ram configured for each one): [root at virtfed cluster]# free total used free shared buffers cached Mem: 12330828 7233912 5096916 0 1567600 245032 -/+ buffers/cache: 5421280 6909548 Swap: 4194296 15596 4178700 From anders+fedora-virt at trudheim.co.uk Mon Nov 30 20:23:18 2009 From: anders+fedora-virt at trudheim.co.uk (Anders Rayner-Karlsson) Date: Mon, 30 Nov 2009 21:23:18 +0100 Subject: [fedora-virt] some vm die in F12 In-Reply-To: <561c252c0911300658k2059e39aoa0e353f2aece377c@mail.gmail.com> References: <561c252c0911300600r3edd52a3h175c81ff25e30911@mail.gmail.com> <4B13D8D3.4020805@redhat.com> <561c252c0911300648i5091d17fn1e29eb541f7ecc64@mail.gmail.com> <561c252c0911300658k2059e39aoa0e353f2aece377c@mail.gmail.com> Message-ID: <20091130202318.GJ7408@shuttle> * Gianluca Cecchi [20091130 15:59]: > On Mon, Nov 30, 2009 at 3:48 PM, Gianluca Cecchi [snip] > I wrote about host having 10Gb of ram. > Actually it has 12Gb of ram (as seen btw from the output of the "free" command). > And after starting the other two guests (3,5Gb of ram configured for each one): > [root at virtfed cluster]# free > total used free shared buffers cached > Mem: 12330828 7233912 5096916 0 1567600 245032 > -/+ buffers/cache: 5421280 6909548 > Swap: 4194296 15596 4178700 It would not perchance be BZ#540330 you're hitting would it? Worth checking what the size is of libvirtd periodically perhaps, if nothing else, to rule out that it's this BZ. Just my ?2 :) /Anders