From michael.coffman at avagotech.com Mon Apr 2 13:58:46 2012 From: michael.coffman at avagotech.com (Michael Coffman) Date: Mon, 2 Apr 2012 07:58:46 -0600 Subject: [rhelv6-list] chkfontpath missing Message-ID: chkfontpath is missing from rhel6. Is there a replacement command on 6 for this functionality? xfs also appears to be missing. any replacement/alternative? Thanks. -MichaelC -------------- next part -------------- An HTML attachment was scrubbed... URL: From rprice at redhat.com Mon Apr 2 14:29:24 2012 From: rprice at redhat.com (Robin Price II) Date: Mon, 02 Apr 2012 09:29:24 -0500 Subject: [rhelv6-list] chkfontpath missing In-Reply-To: References: Message-ID: <4F79B7C4.1030702@redhat.com> Michael, chkfontpath appears to be in atrpms: http://atrpms.net/ [rprice at T520 ~]$ grep fontpath yum-list.txt chkfontpath.x86_64 1.10.1-2.el6 atrpms XFS is an add-on: http://www.redhat.com/products/enterprise-linux-add-ons/file-systems/ Hope this helps. ~rp On 04/02/2012 08:58 AM, Michael Coffman wrote: > > chkfontpath is missing from rhel6. Is there a replacement command on > 6 for this functionality? > > xfs also appears to be missing. any replacement/alternative? > > Thanks. > -MichaelC > > > _______________________________________________ > rhelv6-list mailing list > rhelv6-list at redhat.com > https://www.redhat.com/mailman/listinfo/rhelv6-list -- +-----------------------------[ robin at redhat.com ]----+ | Robin Price II - RHCE,RHCDS,RHCVA | | Inside Solutions Architect | | Red Hat, Inc. | | w: +1 (919) 754 4412 | | c: +1 (252) 474 3525 | | | +---------[ http://people.redhat.com/rprice ]---------+ From michael.coffman at avagotech.com Mon Apr 2 14:34:48 2012 From: michael.coffman at avagotech.com (Michael Coffman) Date: Mon, 2 Apr 2012 08:34:48 -0600 Subject: [rhelv6-list] chkfontpath missing In-Reply-To: <4F79B7C4.1030702@redhat.com> References: <4F79B7C4.1030702@redhat.com> Message-ID: On Mon, Apr 2, 2012 at 8:29 AM, Robin Price II wrote: > Michael, > > chkfontpath appears to be in atrpms: > http://atrpms.net/ > > I had found this but still needed xfs. > [rprice at T520 ~]$ grep fontpath yum-list.txt > chkfontpath.x86_64 1.10.1-2.el6 atrpms > > XFS is an add-on: > http://www.redhat.com/**products/enterprise-linux-add-**ons/file-systems/ > > Hope this helps. > > Didn't even look there, thanks for the pointer. Is there another way to deal with this? I am happy to change if there is an alternative that provides the same or similar functionality. I just don't see anything beyond the fontconfig package. > ~rp > > > On 04/02/2012 08:58 AM, Michael Coffman wrote: > >> >> chkfontpath is missing from rhel6. Is there a replacement command on >> 6 for this functionality? >> >> xfs also appears to be missing. any replacement/alternative? >> >> Thanks. >> -MichaelC >> >> >> ______________________________**_________________ >> rhelv6-list mailing list >> rhelv6-list at redhat.com >> https://www.redhat.com/**mailman/listinfo/rhelv6-list >> > > > -- > +-----------------------------**[ robin at redhat.com ]----+ > | Robin Price II - RHCE,RHCDS,RHCVA | > | Inside Solutions Architect | > | Red Hat, Inc. | > | w: +1 (919) 754 4412 | > | c: +1 (252) 474 3525 | > | | > +---------[ http://people.redhat.com/**rprice]---------+ > > ______________________________**_________________ > rhelv6-list mailing list > rhelv6-list at redhat.com > https://www.redhat.com/**mailman/listinfo/rhelv6-list > -------------- next part -------------- An HTML attachment was scrubbed... URL: From markmont at umich.edu Mon Apr 2 14:35:00 2012 From: markmont at umich.edu (Mark Montague) Date: Mon, 02 Apr 2012 10:35:00 -0400 Subject: [rhelv6-list] chkfontpath missing In-Reply-To: References: Message-ID: <4F79B914.70801@umich.edu> On April 2, 2012 9:58 , Michael Coffman wrote: > chkfontpath is missing from rhel6. Is there a replacement command > on 6 for this functionality? Yes, create a symlink in /etc/X11/fontpath.d For an explanation of why chkfontpath and xfs have gone away, plus what people should do instead, see https://fedoraproject.org/wiki/Releases/FeatureNoMoreXFS > xfs also appears to be missing. any replacement/alternative? It "should not" be necessary, according to what I understand from what I have read. If you have a very special situation and need to run it, you could try back-porting https://admin.fedoraproject.org/pkgdb/acls/name/xorg-x11-xfs There are also third-party repositories that offer xfs packages for RHEL6, but how trustworthy they are I don't know. -- Mark Montague LSA Research Systems Group University of Michigan markmont at umich.edu From michael.coffman at avagotech.com Mon Apr 2 14:36:50 2012 From: michael.coffman at avagotech.com (Michael Coffman) Date: Mon, 2 Apr 2012 08:36:50 -0600 Subject: [rhelv6-list] chkfontpath missing In-Reply-To: References: <4F79B7C4.1030702@redhat.com> Message-ID: On Mon, Apr 2, 2012 at 8:34 AM, Michael Coffman < michael.coffman at avagotech.com> wrote: > On Mon, Apr 2, 2012 at 8:29 AM, Robin Price II wrote: > >> Michael, >> >> chkfontpath appears to be in atrpms: >> http://atrpms.net/ >> >> > I had found this but still needed xfs. > > >> [rprice at T520 ~]$ grep fontpath yum-list.txt >> chkfontpath.x86_64 1.10.1-2.el6 atrpms >> >> XFS is an add-on: >> http://www.redhat.com/**products/enterprise-linux-add-**ons/file-systems/ >> >> Hope this helps. >> >> > Didn't even look there, thanks for the pointer. > > Woops. The is xfs as in file system, not font server :( > Is there another way to deal with this? I am happy to change if there is > an alternative that provides the same or similar functionality. I just > don't see anything beyond the fontconfig package. > > >> ~rp >> >> >> On 04/02/2012 08:58 AM, Michael Coffman wrote: >> >>> >>> chkfontpath is missing from rhel6. Is there a replacement command on >>> 6 for this functionality? >>> >>> xfs also appears to be missing. any replacement/alternative? >>> >>> Thanks. >>> -MichaelC >>> >>> >>> ______________________________**_________________ >>> rhelv6-list mailing list >>> rhelv6-list at redhat.com >>> https://www.redhat.com/**mailman/listinfo/rhelv6-list >>> >> >> >> -- >> +-----------------------------**[ robin at redhat.com ]----+ >> | Robin Price II - RHCE,RHCDS,RHCVA | >> | Inside Solutions Architect | >> | Red Hat, Inc. | >> | w: +1 (919) 754 4412 | >> | c: +1 (252) 474 3525 | >> | | >> +---------[ http://people.redhat.com/**rprice]---------+ >> >> ______________________________**_________________ >> rhelv6-list mailing list >> rhelv6-list at redhat.com >> https://www.redhat.com/**mailman/listinfo/rhelv6-list >> > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From jsbillin at umich.edu Mon Apr 2 14:41:32 2012 From: jsbillin at umich.edu (Jonathan Billings) Date: Mon, 2 Apr 2012 10:41:32 -0400 Subject: [rhelv6-list] chkfontpath missing In-Reply-To: <4F79B7C4.1030702@redhat.com> References: <4F79B7C4.1030702@redhat.com> Message-ID: On Mon, Apr 2, 2012 at 10:29 AM, Robin Price II wrote: > XFS is an add-on: > http://www.redhat.com/products/enterprise-linux-add-ons/file-systems/ The original question was about the X Font Server, not the XFS Filesystem. They're both called 'xfs' but refer to completely different things. -- Jonathan Billings College of Engineering - CAEN - Unix and Linux Support From bill at magicdigits.com Mon Apr 2 14:41:33 2012 From: bill at magicdigits.com (Bill Watson) Date: Mon, 2 Apr 2012 07:41:33 -0700 Subject: [rhelv6-list] Red Hat 6.2 update woes Message-ID: <049301cd10de$b30de9b0$1929bd10$@com> Howdy, I installed RHEL 6.0, then "yum update'ed to apply all patches. This fouled my new kernel (2.6.32-220.7.1.el6.x86_64) with the following: [drm:drm edid_block_valid] *ERROR* EDID checksum is invalid, remainder is 88 (repeated 3 more times) [drm] nouveau 0000:06:00:0 DDC responded, but no EDID for VGA-1 Kernel panic I have a GdForce 8400GS IO Express 2.0 graphics card w/DDR3 I got the system back up by booting the original kernel (2.6.32-71.el6.x86_64). My hardware is the GeGorce 8400GS, which appears to have some google infamy, but nothing this recently. There is no mention of RedHat on the video card website wvga.com in the driver section. Is there a glimmer of hope that this card can work with RHEL 6.2? Upon further research, I tried booting the kernel, and at the choose which kernel screen, added: This removed display of the EDID checksum errors, but still kernel panic'ed with the dreaded: kernel panic - not syncing: Attempted to kill init! All the helpful hinters said to check /var/log/* and this seemed to be unhelpful since "not syncing" seems to also mean "not writing useful stuff onto the log files". It is possible that though the following Adaptec 6805 card can read enough to boot, it doesn't see the filesystems when the kernel takes over. Also I have found that the Adaptec 6805 RAID SAS/SATA card is tough to install. I did find a path where installing RHEL 6.0 does work. I have found no path to install 6.2 directly. For quite a while I have found Adaptec to be basically native to Red Hat, but this card is hostile to install. As far as that goes, I got RHEL 5.5 working, but not 5.8 for installation. I saw a mention of kmod-aacraid-1.1.7-3.el6.elrepo.x86_64.rpm over on Dag's http://elrepo.org/linux/elrepo/el6/x86_64/RPMS/ that sounded really tempting to install, but I don't know if it would blow up my working original kernel or not. Until I am confident of a direction, I'd kind of like to keep this kernel working so I can continue to try this and that. Thanks all in advance, Bill Watson bill at magicdigits.com From rprice at redhat.com Mon Apr 2 15:01:00 2012 From: rprice at redhat.com (Robin Price II) Date: Mon, 02 Apr 2012 10:01:00 -0500 Subject: [rhelv6-list] chkfontpath missing In-Reply-To: References: <4F79B7C4.1030702@redhat.com> Message-ID: <4F79BF2C.2030605@redhat.com> On 04/02/2012 09:41 AM, Jonathan Billings wrote: > On Mon, Apr 2, 2012 at 10:29 AM, Robin Price II wrote: >> XFS is an add-on: >> http://www.redhat.com/products/enterprise-linux-add-ons/file-systems/ > > The original question was about the X Font Server, not the XFS > Filesystem. They're both called 'xfs' but refer to completely > different things. > I blame it on it being Monday. derp. :) ~rp -- +-----------------------------[ robin at redhat.com ]----+ | Robin Price II - RHCE,RHCDS,RHCVA | | Inside Solutions Architect | | Red Hat, Inc. | | w: +1 (919) 754 4412 | | c: +1 (252) 474 3525 | | @robinpriceii | +---------[ http://people.redhat.com/rprice ]---------+ From rprice at redhat.com Mon Apr 2 15:01:33 2012 From: rprice at redhat.com (Robin Price II) Date: Mon, 02 Apr 2012 10:01:33 -0500 Subject: [rhelv6-list] chkfontpath missing In-Reply-To: References: <4F79B7C4.1030702@redhat.com> Message-ID: <4F79BF4D.6010108@redhat.com> LOL Wow. I really need my coffee. So sorry for the confusion. :( ~rp On 04/02/2012 09:36 AM, Michael Coffman wrote: > On Mon, Apr 2, 2012 at 8:34 AM, Michael Coffman > > > wrote: > > On Mon, Apr 2, 2012 at 8:29 AM, Robin Price II > wrote: > > Michael, > > chkfontpath appears to be in atrpms: > http://atrpms.net/ > > > I had found this but still needed xfs. > > [rprice at T520 ~]$ grep fontpath yum-list.txt > chkfontpath.x86_64 1.10.1-2.el6 atrpms > > XFS is an add-on: > http://www.redhat.com/__products/enterprise-linux-add-__ons/file-systems/ > > > Hope this helps. > > > Didn't even look there, thanks for the pointer. > > > Woops. The is xfs as in file system, not font server :( > > Is there another way to deal with this? I am happy to change if > there is an alternative that provides the same or similar > functionality. I just don't see anything beyond the fontconfig package. > > ~rp > > > On 04/02/2012 08:58 AM, Michael Coffman wrote: > > > chkfontpath is missing from rhel6. Is there a replacement > command on > 6 for this functionality? > > xfs also appears to be missing. any replacement/alternative? > > Thanks. > -MichaelC > > > _________________________________________________ > rhelv6-list mailing list > rhelv6-list at redhat.com > https://www.redhat.com/__mailman/listinfo/rhelv6-list > > > > > -- > +-----------------------------__[ robin at redhat.com > ]----+ > | Robin Price II - RHCE,RHCDS,RHCVA | > | Inside Solutions Architect | > | Red Hat, Inc. | > | w: +1 (919) 754 4412 > | > | c: +1 (252) 474 3525 > | > | | > +---------[ http://people.redhat.com/__rprice > ]---------+ > > _________________________________________________ > rhelv6-list mailing list > rhelv6-list at redhat.com > https://www.redhat.com/__mailman/listinfo/rhelv6-list > > > > > > > _______________________________________________ > rhelv6-list mailing list > rhelv6-list at redhat.com > https://www.redhat.com/mailman/listinfo/rhelv6-list -- +-----------------------------[ robin at redhat.com ]----+ | Robin Price II - RHCE,RHCDS,RHCVA | | Inside Solutions Architect | | Red Hat, Inc. | | w: +1 (919) 754 4412 | | c: +1 (252) 474 3525 | | @robinpriceii | +---------[ http://people.redhat.com/rprice ]---------+ From amyagi at gmail.com Mon Apr 2 15:10:19 2012 From: amyagi at gmail.com (Akemi Yagi) Date: Mon, 2 Apr 2012 08:10:19 -0700 Subject: [rhelv6-list] Red Hat 6.2 update woes In-Reply-To: <049301cd10de$b30de9b0$1929bd10$@com> References: <049301cd10de$b30de9b0$1929bd10$@com> Message-ID: On Mon, Apr 2, 2012 at 7:41 AM, Bill Watson wrote: > Also I have found that the Adaptec 6805 RAID SAS/SATA card is tough to > install. I did find a path where installing RHEL 6.0 does work. I have found > no path to install 6.2 directly. For quite a while I have found Adaptec to > be basically native to Red Hat, but this card is hostile to install. ?As far > as that goes, I got RHEL 5.5 working, but not 5.8 for installation. > > I saw a mention of ?kmod-aacraid-1.1.7-3.el6.elrepo.x86_64.rpm ?over on > Dag's http://elrepo.org/linux/elrepo/el6/x86_64/RPMS/ that sounded really > tempting to install, but I don't know if it would blow up my working > original kernel or not. > Until I am confident of a direction, I'd kind of like to keep this kernel > working so I can continue to try this and that. To install RHEL, you'd need a driver disk for the Adaptec card. ELRepo (Phil Schaffner) provides one for the the 6000 series here: http://www.elrepo.org/people/pschaff/dud/ as DUP_el6_i686_20111230.iso.gz (32-bit) and DUP_el6_x86_64_20111230.iso.gz (64-bit). More details can be found here: http://elrepo.org/bugs/view.php?id=206 Will be great if you can test it. Akemi From bill at magicdigits.com Mon Apr 2 15:34:15 2012 From: bill at magicdigits.com (Bill Watson) Date: Mon, 2 Apr 2012 08:34:15 -0700 Subject: [rhelv6-list] Red Hat 6.2 update woes In-Reply-To: References: <049301cd10de$b30de9b0$1929bd10$@com> Message-ID: <057101cd10e6$1664ce00$432e6a00$@com> > I saw a mention of ?kmod-aacraid-1.1.7-3.el6.elrepo.x86_64.rpm ?over > on Dag's http://elrepo.org/linux/elrepo/el6/x86_64/RPMS/ that sounded > really tempting to install, but I don't know if it would blow up my > working original kernel or not. > Until I am confident of a direction, I'd kind of like to keep this > kernel working so I can continue to try this and that. To install RHEL, you'd need a driver disk for the Adaptec card. ELRepo (Phil Schaffner) provides one for the the 6000 series here: http://www.elrepo.org/people/pschaff/dud/ as DUP_el6_i686_20111230.iso.gz (32-bit) and DUP_el6_x86_64_20111230.iso.gz (64-bit). More details can be found here: http://elrepo.org/bugs/view.php?id=206 Will be great if you can test it. Akemi -------------------------------------------------- I downloaded this iso.gz and found it had the kmod-aacraid-1.1.7-1. rpm, but interestingly it had a different number of bytes than the same named rpm from elrepo.org. There is no clear indication, but I currently presume that the 7-3 rpm is newer thanthe 7-1 one and hopefully this makes is a better choice. Bill From amyagi at gmail.com Mon Apr 2 16:16:03 2012 From: amyagi at gmail.com (Akemi Yagi) Date: Mon, 2 Apr 2012 09:16:03 -0700 Subject: [rhelv6-list] Red Hat 6.2 update woes In-Reply-To: <057101cd10e6$1664ce00$432e6a00$@com> References: <049301cd10de$b30de9b0$1929bd10$@com> <057101cd10e6$1664ce00$432e6a00$@com> Message-ID: On Mon, Apr 2, 2012 at 8:34 AM, Bill Watson wrote: >> I saw a mention of ?kmod-aacraid-1.1.7-3.el6.elrepo.x86_64.rpm ?over >> on Dag's http://elrepo.org/linux/elrepo/el6/x86_64/RPMS/ that sounded >> really tempting to install, but I don't know if it would blow up my >> working original kernel or not. >> Until I am confident of a direction, I'd kind of like to keep this >> kernel working so I can continue to try this and that. > > To install RHEL, you'd need a driver disk for the Adaptec card. ELRepo (Phil > Schaffner) provides one for the the 6000 series here: > > http://www.elrepo.org/people/pschaff/dud/ > > as DUP_el6_i686_20111230.iso.gz (32-bit) and DUP_el6_x86_64_20111230.iso.gz > (64-bit). More details can be found > here: > > http://elrepo.org/bugs/view.php?id=206 > > Will be great if you can test it. > > Akemi > -------------------------------------------------- > I downloaded this iso.gz and found it had the kmod-aacraid-1.1.7-1. rpm, but > interestingly it had a different number of bytes than the same named rpm > from elrepo.org. There is no clear indication, but I currently presume that > the 7-3 rpm is newer thanthe 7-1 one and hopefully this makes is a better > choice. > Bill I'm sure Phil (ELRepo) will update the driver disk file with the latest aacraid driver. I think the file name should contain the driver version rather than (or in addition to) the date. :) Akemi From Philip.R.Schaffner at nasa.gov Mon Apr 2 22:01:56 2012 From: Philip.R.Schaffner at nasa.gov (Phil Schaffner) Date: Mon, 2 Apr 2012 18:01:56 -0400 Subject: [rhelv6-list] Red Hat 6.2 update woes In-Reply-To: References: <049301cd10de$b30de9b0$1929bd10$@com> <057101cd10e6$1664ce00$432e6a00$@com> Message-ID: <4F7A21D4.5040406@NASA.gov> Akemi Yagi wrote on 04/02/2012 12:16 PM: > I'm sure Phil (ELRepo) will update the driver disk file with the > latest aacraid driver. I think the file name should contain the driver > version rather than (or in addition to) the date. :) New ISO images have been built and uploaded: DUP_el6_i686_20120402.iso.gz DUP_el6_x86_64_20120402.iso.gz The images contain other drivers in addition to aacraid, so labeling with a single version would be misleading. Phil From bill at magicdigits.com Tue Apr 3 01:34:57 2012 From: bill at magicdigits.com (Bill Watson) Date: Mon, 2 Apr 2012 18:34:57 -0700 Subject: [rhelv6-list] Red Hat 6.2 update woes In-Reply-To: <4F7A21D4.5040406@NASA.gov> References: <049301cd10de$b30de9b0$1929bd10$@com> <057101cd10e6$1664ce00$432e6a00$@com> <4F7A21D4.5040406@NASA.gov> Message-ID: <0e2001cd1139$fa9e2a50$efda7ef0$@com> Akemi Yagi wrote on 04/02/2012 12:16 PM: > I'm sure Phil (ELRepo) will update the driver disk file with the > latest aacraid driver. I think the file name should contain the driver > version rather than (or in addition to) the date. :) New ISO images have been built and uploaded: DUP_el6_i686_20120402.iso.gz DUP_el6_x86_64_20120402.iso.gz The images contain other drivers in addition to aacraid, so labeling with a single version would be misleading. Phil ------------------------------------ [:solved:] It works, it works!!! I already used the 7-3 version from elrepo.org when I saw your email. I have documented my path to success at the following web link: http://www.magicdigits.com/adaptec/howto.txt (2.6k) also that folder has the initial installation floppy image as well as the post installation rpm that appear both to be needed for this controller to work. What is documented encompases what I've read on many sites to create a soup to nuts installation guideline. I have not seen any single site with everything it takes for success, which compelled me to make this one (my first of its kind). Of course, now that this is documented, tomorrow Red Hat will come out with these drivers native in the kernel. =)~ The reason using this card was important to me is that it is one of the few hybrid RAID controllers where I was able to mirror a 128mb (SLC) ssd to a 2tb sata, then mirror another pair, and then mirror the remaining balance of the 2tb disks to each other. My world fits on 128mb with room to spare, so this makes a hyper-high speed system with lots of long term online storage. Thanks to all who blazed their portion of this path for me to use as my pallette. Bill Watson bill at magicdigits.com From Philip.R.Schaffner at nasa.gov Tue Apr 3 11:56:16 2012 From: Philip.R.Schaffner at nasa.gov (Phil Schaffner) Date: Tue, 3 Apr 2012 07:56:16 -0400 Subject: [rhelv6-list] Red Hat 6.2 update woes In-Reply-To: <0e2001cd1139$fa9e2a50$efda7ef0$@com> References: <049301cd10de$b30de9b0$1929bd10$@com> <057101cd10e6$1664ce00$432e6a00$@com> <4F7A21D4.5040406@NASA.gov> <0e2001cd1139$fa9e2a50$efda7ef0$@com> Message-ID: <4F7AE560.7080902@NASA.gov> Bill Watson wrote on 04/02/2012 09:34 PM: > Akemi Yagi wrote on 04/02/2012 12:16 PM: >> I'm sure Phil (ELRepo) will update the driver disk file with the >> latest aacraid driver. I think the file name should contain the driver >> version rather than (or in addition to) the date. :) > New ISO images have been built and uploaded: > DUP_el6_i686_20120402.iso.gz > DUP_el6_x86_64_20120402.iso.gz > The images contain other drivers in addition to aacraid, so labeling with a > single version would be misleading. > Phil > ------------------------------------ > [:solved:] It works, it works!!! > > I already used the 7-3 version from elrepo.org when I saw your email. > Glad to hear you have resolved your issue. It would still be interesting to know if the driver update disk ISO worked, as that should make the whole process a lot more streamlined. Phil From bill at magicdigits.com Tue Apr 3 15:33:23 2012 From: bill at magicdigits.com (Bill Watson) Date: Tue, 3 Apr 2012 08:33:23 -0700 Subject: [rhelv6-list] Red Hat 6.2 update woes In-Reply-To: <4F7AE560.7080902@NASA.gov> References: <049301cd10de$b30de9b0$1929bd10$@com> <057101cd10e6$1664ce00$432e6a00$@com> <4F7A21D4.5040406@NASA.gov> <0e2001cd1139$fa9e2a50$efda7ef0$@com> <4F7AE560.7080902@NASA.gov> Message-ID: <034b01cd11af$2829f600$787de200$@com> > [:solved:] It works, it works!!! > > I already used the 7-3 version from elrepo.org when I saw your email. > Glad to hear you have resolved your issue. It would still be interesting to know if the driver update disk ISO worked, as that should make the whole process a lot more streamlined. Phil______________________________________________ I do not see how this update disk would, could, or should work any different than the elrepo rpm. Without a doubt the 7-1 version when loaded onto a usb stick does not handle a RHEL 6.0 or 6.2 bare metal install. If you / yall want to write a test procedure that you wish me to follow, I'd be happy to help. Bill Watson bill at magicdigits.com www.magicdigits.com/adaptec From jas at cse.yorku.ca Tue Apr 3 16:31:58 2012 From: jas at cse.yorku.ca (Jason Keltz) Date: Tue, 03 Apr 2012 12:31:58 -0400 Subject: [rhelv6-list] startup time virt-manager In-Reply-To: <1332811954.3772.124.camel@kauai> References: <72F8605A4AEE17448F28CEF3F4E1925F819B4D90EA@MSGRTPCCRC2WIN.DMN1.FMR.COM> <72F8605A4AEE17448F28CEF3F4E1925F819B573236@MSGRTPCCRC2WIN.DMN1.FMR.COM> , <72F8605A4AEE17448F28CEF3F4E1925F819B70C6C0@MSGRTPCCRC2WIN.DMN1.FMR.COM> <1D241511770E2F4BA89AFD224EDD527117B81BE4@G9W0737.americas.hpqcorp.net> <1332804057.3772.65.camel@kauai> <1D241511770E2F4BA89AFD224EDD527117B81C3C@G9W0737.americas.hpqcorp.net> <1332811954.3772.124.camel@kauai> Message-ID: <4F7B25FE.3030701@cse.yorku.ca> Hi. I'm wondering what I should expect the delay to be when starting virt-manager connecting to two hosts like this: host1 # virt-manager -c qemu+ssh://virt1/system -c qemu+ssh://virt2/system There is 1 VM running on each of virt1 and virt2. The delay is approximately 20 seconds in total. It is more than I suspect it should take, but I don't see any errors, it seems to connect, and everything seems to be speedy afterwards. If this isn't normal, any idea on how to debug the delay? I've tried running virt-manager with the --debug option, but it doesn't generate any output during the connection... Jason. From prashanth.sundaram at laurioncap.com Tue Apr 3 20:16:58 2012 From: prashanth.sundaram at laurioncap.com (Prashanth Sundaram) Date: Tue, 3 Apr 2012 16:16:58 -0400 Subject: [rhelv6-list] Seeing only ONE session/device instead of TWO Message-ID: <9B3BA590E4DCC547A60DD2B8E78B4F1F011C09E3@exchange.laurion.corp> Hello, I am having some trouble with ISCSI to work on RHEL 6.2. It currently is using 1 nic to connect to SAN and shows only 1 session instead of 2. I have configured TWO nics(em3 & em4) to perform multipathing and have verified that both can connect to SAN. Any ideas how to get multipath to work? I only see 1 device instead of 2 in multipath. Please see my config below. Another strange thing is that the iface_name has different MAC address than what ifconfig shows. ------------------------------------------------------------------------ ------------------------------------------------ # iscsiadm -m session tcp: [17] 10.10.100.200:3260,1 iqn.2001-05.com.equallogic:8-cb2b76-511758f4a-f4015609f414f7a1-fs-proj-0 1 # multipath -ll fs-proj-01 (368b7b2acf4581751a1f714f4096sd1f4) dm-7 EQLOGIC,100E-00 size=50G features='0' hwhandler='0' wp=rw `-+- policy='round-robin 0' prio=1 status=active `- 7:0:0:0 sdb 8:16 active ready running # iscsiadm -m discovery 10.10.100.200:3260 via sendtargets # iscsiadm -m iface default tcp,,,, iser iser,,,, bnx2i.18:03:73:f2:fc:b7 bnx2i,18:03:73:f2:fc:b7,,, bnx2i.18:03:73:f2:fc:b9 bnx2i,18:03:73:f2:fc:b9,,, bnx2i.18:03:73:f2:fc:b3 bnx2i,18:03:73:f2:fc:b3,,, bnx2i.18:03:73:f2:fc:b5 bnx2i,18:03:73:f2:fc:b5,,, # cat /etc/iscsi/iscsid.conf discovery.sendtargets.iscsi.MaxRecvDataSegmentLength = 32768 node.conn[0].iscsi.HeaderDigest = None node.conn[0].iscsi.MaxRecvDataSegmentLength = 262144 node.conn[0].iscsi.MaxXmitDataSegmentLength = 0 node.conn[0].timeo.login_timeout = 15 node.conn[0].timeo.logout_timeout = 15 node.conn[0].timeo.noop_out_interval = 5 node.conn[0].timeo.noop_out_timeout = 5 node.leading_login = No node.session.cmds_max = 1024 node.session.err_timeo.abort_timeout = 15 node.session.err_timeo.lu_reset_timeout = 30 node.session.err_timeo.tgt_reset_timeout = 30 node.session.initial_login_retry_max = 8 node.session.iscsi.FastAbort = Yes node.session.iscsi.FirstBurstLength = 262144 node.session.iscsi.ImmediateData = Yes node.session.iscsi.InitialR2T = No node.session.iscsi.MaxBurstLength = 16776192 node.session.nr_sessions = 2 node.session.queue_depth = 128 node.session.timeo.replacement_timeout = 120 node.session.xmit_thread_priority = -20 node.startup = automatic Thanks, Prashanth -------------- next part -------------- An HTML attachment was scrubbed... URL: From pmeyer at themeyerfarm.com Tue Apr 3 21:32:42 2012 From: pmeyer at themeyerfarm.com (Phil Meyer) Date: Tue, 03 Apr 2012 15:32:42 -0600 Subject: [rhelv6-list] startup time virt-manager In-Reply-To: <4F7B25FE.3030701@cse.yorku.ca> References: <72F8605A4AEE17448F28CEF3F4E1925F819B4D90EA@MSGRTPCCRC2WIN.DMN1.FMR.COM> <72F8605A4AEE17448F28CEF3F4E1925F819B573236@MSGRTPCCRC2WIN.DMN1.FMR.COM> , <72F8605A4AEE17448F28CEF3F4E1925F819B70C6C0@MSGRTPCCRC2WIN.DMN1.FMR.COM> <1D241511770E2F4BA89AFD224EDD527117B81BE4@G9W0737.americas.hpqcorp.net> <1332804057.3772.65.camel@kauai> <1D241511770E2F4BA89AFD224EDD527117B81C3C@G9W0737.americas.hpqcorp.net> <1332811954.3772.124.camel@kauai> <4F7B25FE.3030701@cse.yorku.ca> Message-ID: <4F7B6C7A.7010209@themeyerfarm.com> On 04/03/2012 10:31 AM, Jason Keltz wrote: > Hi. > > I'm wondering what I should expect the delay to be when starting > virt-manager connecting to two hosts like this: > > host1 # virt-manager -c qemu+ssh://virt1/system -c > qemu+ssh://virt2/system > > There is 1 VM running on each of virt1 and virt2. > > The delay is approximately 20 seconds in total. It is more than I > suspect it > should take, but I don't see any errors, it seems to connect, and > everything seems > to be speedy afterwards. > > If this isn't normal, any idea on how to debug the delay? I've tried > running > virt-manager with the --debug option, but it doesn't generate any > output during > the connection... Reverse DNS lookup error? That amount of time is typical of that. From pmeyer at themeyerfarm.com Tue Apr 3 21:43:57 2012 From: pmeyer at themeyerfarm.com (Phil Meyer) Date: Tue, 03 Apr 2012 15:43:57 -0600 Subject: [rhelv6-list] Seeing only ONE session/device instead of TWO In-Reply-To: <9B3BA590E4DCC547A60DD2B8E78B4F1F011C09E3@exchange.laurion.corp> References: <9B3BA590E4DCC547A60DD2B8E78B4F1F011C09E3@exchange.laurion.corp> Message-ID: <4F7B6F1D.9030700@themeyerfarm.com> On 04/03/2012 02:16 PM, Prashanth Sundaram wrote: > > Hello, > > I am having some trouble with ISCSI to work on RHEL 6.2. It currently > is using 1 nic to connect to SAN and shows only 1 session instead of > 2. I have configured TWO nics(em3 & em4) to perform multipathing and > have verified that both can connect to SAN. > > Any ideas how to get multipath to work? I only see 1 device instead of > 2 in multipath. Please see my config below. > > Another strange thing is that the iface_name has different MAC address > than what ifconfig shows. > > Two main issues. On the same network, two interfaces need to negotiate ARP independently of each other. If not, a packet can exit one interface, and be sent a response to another interface. That does not work. /etc/sysctl.conf needs to be adjusted for ARP by interface. Clues: # net # net.ipv4.conf.all.arp_filter = 1 # net.ipv4.conf.all.rp_filter = 1 # net.ipv4.conf.default.arp_filter = 1 net.ipv4.conf.all.arp_ignore=1 net.ipv4.conf.all.arp_announce=2 ISCSI needs to be told that there are two interfaces, it does not know it by default. This is done by special files in: /var/lib/iscsi/ifaces I don't have a sample nearby, but that should get you googleing ok. Once that is in place, 'iscsiadm -m node -l' will show two logins for every target. If THAT happens, 'multipath -r' will show two connections per device. Good Luck! From prashanth.sundaram at laurioncap.com Tue Apr 3 22:05:26 2012 From: prashanth.sundaram at laurioncap.com (Prashanth Sundaram) Date: Tue, 3 Apr 2012 18:05:26 -0400 Subject: [rhelv6-list] Seeing only ONE session/device instead of TWO In-Reply-To: <4F7B6F1D.9030700@themeyerfarm.com> References: <9B3BA590E4DCC547A60DD2B8E78B4F1F011C09E3@exchange.laurion.corp> <4F7B6F1D.9030700@themeyerfarm.com> Message-ID: <9B3BA590E4DCC547A60DD2B8E78B4F1F011C09E4@exchange.laurion.corp> Thanks for the response Phil. /etc/sysctl.conf is correct per your description. I see four files in there. What surprises me is that the ifacename, which is a combination of HWaddr and Mfr does not seem to match the mac address reported by the system. See below, the last value is an increment or something. # ls /var/lib/iscsi/ifaces/ bnx2i.18:03:73:f2:fc:b3 bnx2i.18:03:73:f2:fc:b5 bnx2i.18:03:73:f2:fc:b7 bnx2i.18:03:73:f2:fc:b9 # ifconfig | grep 'HWaddr' em1 Link encap:Ethernet HWaddr 18:03:73:F2:FC:B2 em2 Link encap:Ethernet HWaddr 18:03:73:F2:FC:B4 em3 Link encap:Ethernet HWaddr 18:03:73:F2:FC:B6 em4 Link encap:Ethernet HWaddr 18:03:73:F2:FC:B8 Also in the actual file. I tried deleting, but it picks up the same name and values. For example, see the entry # BEGIN RECORD 2.0-872.33.el6 iface.iscsi_ifacename = bnx2i.18:03:73:f2:fc:b7 iface.hwaddress = 18:03:73:f2:fc:b7 iface.transport_name = bnx2i iface.vlan_id = 0 iface.vlan_priority = 0 iface.iface_num = 0 iface.mtu = 0 iface.port = 0 # END RECORD I am referring to this guide. http://docs.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/6/html/Storag e_Administration_Guide/iscsioffloadmain.html My NIC is capable of Offloading ISCSI, but not sure if I should enable it. So If I stick to Software_ISCSI, I should see 2 connections. The interface used by iscsi is called "default" # iscsiadm -m discovery -t st -p 10.10.100.200:3260 -P 1 Target: iqn.2001-05.com.equallogic:8-cb2b76-511758f4a-f4016309f414f7a1-fs-proj-0 1 Portal: 10.10.100.200:3260,1 Iface Name: default Still not sure where the problem lies. -----Original Message----- From: rhelv6-list-bounces at redhat.com [mailto:rhelv6-list-bounces at redhat.com] On Behalf Of Phil Meyer Sent: Tuesday, April 03, 2012 5:44 PM To: rhelv6-list at redhat.com Subject: Re: [rhelv6-list] Seeing only ONE session/device instead of TWO On 04/03/2012 02:16 PM, Prashanth Sundaram wrote: > > Hello, > > I am having some trouble with ISCSI to work on RHEL 6.2. It currently > is using 1 nic to connect to SAN and shows only 1 session instead of > 2. I have configured TWO nics(em3 & em4) to perform multipathing and > have verified that both can connect to SAN. > > Any ideas how to get multipath to work? I only see 1 device instead of > 2 in multipath. Please see my config below. > > Another strange thing is that the iface_name has different MAC address > than what ifconfig shows. > > Two main issues. On the same network, two interfaces need to negotiate ARP independently of each other. If not, a packet can exit one interface, and be sent a response to another interface. That does not work. /etc/sysctl.conf needs to be adjusted for ARP by interface. Clues: # net # net.ipv4.conf.all.arp_filter = 1 # net.ipv4.conf.all.rp_filter = 1 # net.ipv4.conf.default.arp_filter = 1 net.ipv4.conf.all.arp_ignore=1 net.ipv4.conf.all.arp_announce=2 ISCSI needs to be told that there are two interfaces, it does not know it by default. This is done by special files in: /var/lib/iscsi/ifaces I don't have a sample nearby, but that should get you googleing ok. Once that is in place, 'iscsiadm -m node -l' will show two logins for every target. If THAT happens, 'multipath -r' will show two connections per device. Good Luck! _______________________________________________ rhelv6-list mailing list rhelv6-list at redhat.com https://www.redhat.com/mailman/listinfo/rhelv6-list From jas at cse.yorku.ca Tue Apr 3 23:45:43 2012 From: jas at cse.yorku.ca (Jason Keltz) Date: Tue, 03 Apr 2012 19:45:43 -0400 Subject: [rhelv6-list] startup time virt-manager Message-ID: Hi Phil, If that was the case, ssh wouldn't allow passwordless root login and it would complain... I can ssh fine and there is no delay. Any other ideas how I might debug? Both hosts have 4 IPs and they all reverse lookup fine. Thanks.. Jas Sent from Samsung Mobile Phil Meyer wrote: On 04/03/2012 10:31 AM, Jason Keltz wrote: > Hi. > > I'm wondering what I should expect the delay to be when starting > virt-manager connecting to two hosts like this: > > host1 # virt-manager -c qemu+ssh://virt1/system -c > qemu+ssh://virt2/system > > There is 1 VM running on each of virt1 and virt2. > > The delay is approximately 20 seconds in total.? It is more than I > suspect it > should take, but I don't see any errors, it seems to connect, and > everything seems > to be speedy afterwards. > > If this isn't normal, any idea on how to debug the delay?? I've tried > running > virt-manager with the --debug option, but it? doesn't generate any > output during > the connection... Reverse DNS lookup error? That amount of time is typical of that. _______________________________________________ rhelv6-list mailing list rhelv6-list at redhat.com https://www.redhat.com/mailman/listinfo/rhelv6-list -------------- next part -------------- An HTML attachment was scrubbed... URL: From Philip.R.Schaffner at nasa.gov Wed Apr 4 14:32:19 2012 From: Philip.R.Schaffner at nasa.gov (Phil Schaffner) Date: Wed, 4 Apr 2012 10:32:19 -0400 Subject: [rhelv6-list] Red Hat 6.2 update woes In-Reply-To: <034b01cd11af$2829f600$787de200$@com> References: <049301cd10de$b30de9b0$1929bd10$@com> <057101cd10e6$1664ce00$432e6a00$@com> <4F7A21D4.5040406@NASA.gov> <0e2001cd1139$fa9e2a50$efda7ef0$@com> <4F7AE560.7080902@NASA.gov> <034b01cd11af$2829f600$787de200$@com> Message-ID: <4F7C5B73.2020507@NASA.gov> Bill Watson wrote on 04/03/2012 11:33 AM: > I do not see how this update disk would, could, or should work any different > than the elrepo rpm. Without a doubt the 7-1 version when loaded onto a usb > stick does not handle a RHEL 6.0 or 6.2 bare metal install. > > If you / yall want to write a test procedure that you wish me to follow, I'd > be happy to help. The driver disk should support a bare-metal install, and I believe it should not require the pre and post scripts and other magic you used to get the driver to work. It should result in an installed system with the kmod driver already loaded. The test procedure is simply to do an install using the driver disk. Your replies would be easier to follow if your mail client did proper quoting. Phil From bill at magicdigits.com Wed Apr 4 17:21:40 2012 From: bill at magicdigits.com (Bill Watson) Date: Wed, 4 Apr 2012 10:21:40 -0700 Subject: [rhelv6-list] Red Hat 6.2 update woes In-Reply-To: <4F7C5B73.2020507@NASA.gov> References: <049301cd10de$b30de9b0$1929bd10$@com> <057101cd10e6$1664ce00$432e6a00$@com> <4F7A21D4.5040406@NASA.gov> <0e2001cd1139$fa9e2a50$efda7ef0$@com> <4F7AE560.7080902@NASA.gov> <034b01cd11af$2829f600$787de200$@com> <4F7C5B73.2020507@NASA.gov> Message-ID: <017101cd1287$66a0c6b0$33e25410$@com> -----Original Message----- Bill Watson wrote on 04/03/2012 11:33 AM: > I do not see how this update disk would, could, or should work any > different than the elrepo rpm. Without a doubt the 7-1 version when > loaded onto a usb stick does not handle a RHEL 6.0 or 6.2 bare metal install. > > If you / yall want to write a test procedure that you wish me to > follow, I'd be happy to help. The driver disk should support a bare-metal install, and I believe it should not require the pre and post scripts and other magic you used to get the driver to work. It should result in an installed system with the kmod driver already loaded. The test procedure is simply to do an install using the driver disk. Your replies would be easier to follow if your mail client did proper quoting. Phil ---------------------------------------------------------------------- Sorry, I suffer with Outlook 2007 - and given the bottom posting lead I'm following, it's doing the best it can. The latest driver seems to have failed: 1) I loaded both the i686 and x86_64 driver iso's as the only 2 files on a usb memory stick. 2) I booted the RHEL 6.2 Boot disk DVD. yada yada then 3) Install or Upgrade an existing system TAB 4) vmlinuz initrd=initrd.img and I added linux dd yada yada then 5) Do you have a driver disk? Yes 6) where? sda, fd0, sr0? --->sda OK 7) "There are multiple partitions on this device which contain the driver disk image. Which one would you like to use? It gives me a choice of only /dev/sda1 OK 8) Select which driver file DUP_el6_i686_20120402.iso or DUP_el6_x86_64_20120402.iso I chose x86 OK 9) Error Failed to read directory /tmp/drivers/rpms/i386: No such file or directory OK 10) detecting hardware.... waiting for hardware to initialize (delay) 11) No new drivers were found on this driver disk. This may indicate that this disk has already been loaded, or that the drivers it contains don't match your hardware. Would you like to manually selsct the driver, continue anyway, or load another driver disk? [Manually choose] and it loops back to 5) Bill From michael.coffman at avagotech.com Wed Apr 4 17:24:11 2012 From: michael.coffman at avagotech.com (Michael Coffman) Date: Wed, 4 Apr 2012 11:24:11 -0600 Subject: [rhelv6-list] Customizing gdm Message-ID: I am trying to customize the behaviour of gdm. The first thing I want to do is make sure that X does not run with the ' -nolisten tcp ' option. I tried adding: [security] DisallowTCP=false to /etc/gdm/custom.conf and it has no affect. Changing the value in /etc/gdm/gdm.schemas works. Is manually editing this file the best way to do this? Is there a command line interface to this file? Thanks. -- -MichaelC -------------- next part -------------- An HTML attachment was scrubbed... URL: From i.mortimer at uq.edu.au Thu Apr 5 02:12:29 2012 From: i.mortimer at uq.edu.au (Ian Mortimer) Date: Thu, 5 Apr 2012 02:12:29 +0000 Subject: [rhelv6-list] Red Hat 6.2 update woes In-Reply-To: <017101cd1287$66a0c6b0$33e25410$@com> References: <049301cd10de$b30de9b0$1929bd10$@com> <057101cd10e6$1664ce00$432e6a00$@com> <4F7A21D4.5040406@NASA.gov> <0e2001cd1139$fa9e2a50$efda7ef0$@com> <4F7AE560.7080902@NASA.gov> <034b01cd11af$2829f600$787de200$@com> <4F7C5B73.2020507@NASA.gov> <017101cd1287$66a0c6b0$33e25410$@com> Message-ID: On Wed, 2012-04-04 at 10:21 -0700, Bill Watson wrote: > The latest driver seems to have failed: > 1) I loaded both the i686 and x86_64 driver iso's as the only 2 files on a > usb memory stick. Just copying the iso image onto the disk doesn't work. You'll need to use dd like this: dd if=DUP_el6_x86_64_20120402.iso of=/dev/sdc Replace /dev/sdc with the device path for your usb key. -- Ian From bill at magicdigits.com Fri Apr 6 23:29:04 2012 From: bill at magicdigits.com (Bill Watson) Date: Fri, 6 Apr 2012 16:29:04 -0700 Subject: [rhelv6-list] Red Hat 6.2 update woes In-Reply-To: References: <049301cd10de$b30de9b0$1929bd10$@com> <057101cd10e6$1664ce00$432e6a00$@com> <4F7A21D4.5040406@NASA.gov> <0e2001cd1139$fa9e2a50$efda7ef0$@com> <4F7AE560.7080902@NASA.gov> <034b01cd11af$2829f600$787de200$@com> <4F7C5B73.2020507@NASA.gov> <017101cd1287$66a0c6b0$33e25410$@com> Message-ID: <033901cd144d$0e05a260$2a10e720$@com> -----Original Message----- From: rhelv6-list-bounces at redhat.com [mailto:rhelv6-list-bounces at redhat.com] On Behalf Of Ian Mortimer Sent: Wednesday, April 04, 2012 7:12 PM To: rhelv6-list at redhat.com Subject: Re: [rhelv6-list] Red Hat 6.2 update woes On Wed, 2012-04-04 at 10:21 -0700, Bill Watson wrote: > The latest driver seems to have failed: > 1) I loaded both the i686 and x86_64 driver iso's as the only 2 files > on a usb memory stick. Just copying the iso image onto the disk doesn't work. You'll need to use dd like this: dd if=DUP_el6_x86_64_20120402.iso of=/dev/sdc Replace /dev/sdc with the device path for your usb key. -- Ian Thank you Ian. I have now dd'ed as suggested above. My destination happened to be of=/dev/sdd1. RHEL6.2 Boot disk does not seem to work with this iso file. My *new* dialog was: 1)Boot the 6.2 Boot cd and TAB to choose options to install new system 2)Added linux dd to the default installation string 3)Do you have a driver disk? Yes 4)Where to load from? sda 5)Multiple partitions.. (note only one partition shown, so not sure why it whines here) choose /dev/sda1 6)"Select the file which is your driver disk image" is the prompt. There are NO disk images on this image of this iso file, only rpm's and such. Just for fun, I tried to select pretty much every available file on the USB stick that the dialog showed, but all said: "Failed to load driver disk from file" Either the test still fails, or I need further guidance. The 6.2 Boot CD "felt" like it prefers iso files to be stored as files on the USB stick instead of as a partition, but that is only a guess on my part. This is *implied by the dialog phrase that starts 6) above. Bill Watson bill at magicdigits.com From michael.coffman at avagotech.com Tue Apr 10 15:45:48 2012 From: michael.coffman at avagotech.com (Michael Coffman) Date: Tue, 10 Apr 2012 09:45:48 -0600 Subject: [rhelv6-list] rpm output changed Message-ID: I would like to be able to programatically determine the macrofiles that rpm is using. I used to be able to easily get this from rpm, but not on rhel6. On rhel5 $ rpm --version RPM version 4.4.2.3 $ rpm -v --showrc |grep macrofiles macrofiles : /usr/lib/rpm/macros:/usr/lib/rpm/ia32e-linux/macros:/usr/lib/rpm/redhat/macros:/etc/rpm/macros.*:/etc/rpm/macros:/etc/rpm/ia32e-linux/macros:~/.rpmmacros On rhel6: $ rpm --version RPM version 4.8.0 $ rpm -v --showrc |grep macrofiles macrofiles : (not set) Is there some place I can depend on to be available on rhel6 that will have the defaults similar to what was shown on rhel5. Thanks. -- -MichaelC -------------- next part -------------- An HTML attachment was scrubbed... URL: From phil at elrepo.org Tue Apr 10 17:31:34 2012 From: phil at elrepo.org (Phil Perry) Date: Tue, 10 Apr 2012 18:31:34 +0100 Subject: [rhelv6-list] rpm output changed In-Reply-To: References: Message-ID: <4F846E76.2020604@elrepo.org> On 10/04/12 16:45, Michael Coffman wrote: > I would like to be able to programatically determine the macrofiles that > rpm is using. I used to be able to easily get this from rpm, but not on > rhel6. > > On rhel5 > $ rpm --version > RPM version 4.4.2.3 > $ rpm -v --showrc |grep macrofiles > macrofiles : > /usr/lib/rpm/macros:/usr/lib/rpm/ia32e-linux/macros:/usr/lib/rpm/redhat/macros:/etc/rpm/macros.*:/etc/rpm/macros:/etc/rpm/ia32e-linux/macros:~/.rpmmacros > > On rhel6: > $ rpm --version > RPM version 4.8.0 > $ rpm -v --showrc |grep macrofiles > macrofiles : (not set) > > Is there some place I can depend on to be available on rhel6 that will have > the defaults similar to what was shown on rhel5. > > Thanks. > Do you have redhat-rpm-config installed? From michael.coffman at avagotech.com Tue Apr 10 18:42:45 2012 From: michael.coffman at avagotech.com (Michael Coffman) Date: Tue, 10 Apr 2012 12:42:45 -0600 Subject: [rhelv6-list] rpm output changed In-Reply-To: <4F846E76.2020604@elrepo.org> References: <4F846E76.2020604@elrepo.org> Message-ID: On Tue, Apr 10, 2012 at 11:31 AM, Phil Perry wrote: > On 10/04/12 16:45, Michael Coffman wrote: > >> I would like to be able to programatically determine the macrofiles that >> rpm is using. I used to be able to easily get this from rpm, but not >> on >> rhel6. >> > > On rhel5 >> $ rpm --version >> RPM version 4.4.2.3 >> $ rpm -v --showrc |grep macrofiles >> macrofiles : >> /usr/lib/rpm/macros:/usr/lib/**rpm/ia32e-linux/macros:/usr/** >> lib/rpm/redhat/macros:/etc/**rpm/macros.*:/etc/rpm/macros:/** >> etc/rpm/ia32e-linux/macros:~/.**rpmmacros >> >> On rhel6: >> $ rpm --version >> RPM version 4.8.0 >> $ rpm -v --showrc |grep macrofiles >> macrofiles : (not set) >> >> Is there some place I can depend on to be available on rhel6 that will >> have >> the defaults similar to what was shown on rhel5. >> >> Thanks. >> >> > Do you have redhat-rpm-config installed? > > Nope. That was what I needed. Thanks! > ______________________________**_________________ > rhelv6-list mailing list > rhelv6-list at redhat.com > https://www.redhat.com/**mailman/listinfo/rhelv6-list > -- -MichaelC -------------- next part -------------- An HTML attachment was scrubbed... URL: From rhelv6-list at redhat.com Tue Apr 24 15:48:54 2012 From: rhelv6-list at redhat.com (Red Hat Enterprise Linux 6 (Santiago) discussion mailing-list) Date: Tue, 24 Apr 2012 11:48:54 -0400 Subject: [rhelv6-list] Red Hat Enterprise Linux 6.3 Beta Announcement Message-ID: <4F96CB66.4030409@redhat.com> Today, Red Hat announces the availability of the public Beta of Red Hat Enterprise Linux 6.3. With this Beta, Red Hat provides early access to the third minor release of Red Hat Enterprise Linux 6, which was initially announced in November 2010. This minor release demonstrates the ongoing effort of Red Hat to innovate with Red Hat Enterprise Linux while maintaining the overall application and platform compatibility across the life cycle. As recently announced here, http://www.redhat.com/about/news/press-archive/2012/1/red-hat-enterprise-linux-stability-drives-demand-for-more-flexibility-in-long-term-operating-system-deployments the Red Hat Enterprise Linux 6 life cycle is now extended to 10 years. The Beta of Red Hat Enterprise Linux 6.3 showcases both enhancements and brand new features in many areas including virtualization, scalability, file systems, storage, security, identity management, and hardware enablement. Although there are nearly 50 new and enhanced features in Red Hat Enterprise Linux 6.3, here are among the most noteworthy: Virtualization -------------- * A new tool called Virt-P2V that facilitates the conversion of physical Windows or Red Hat Enterprise Linux systems into virtual images to be deployed as KVM guests inside Red Hat Enterprise Linux or Red Hat Enterprise Virtualization. * Stronger compliance with Payment Card Industry Data Security Standards (PCI-DSS) standards, including the ability to perform secure wipes of virtual machine disks. * The ability to perform live volume resizing, improving overall availability of virtualized guests. Scalability ----------- * The maximum number of virtual CPUs (vCPUs) has been increased from 64 to 160, which lets you run larger CPU-intensive workloads on the Red Hat Enterprise Linux platform. VMware ESX 5.0 currently supported 32 vCPUs. * The maximum supported memory configuration for KVM guests has been increased from 512GB to 2TB. File Systems ------------ * GFS2 enhancements that create faster read-write capabilities for specific use cases. * Support of O_Direct in FUSE (Filesystem in User Space), which can provide improved performance for certain workloads. * Simplified configuration and administration for the file system. Integration of automount capability with SSSD (System Security Services Daemon) provides centralized management of configuration data and the ability to improve performance through caching and load balancing. (This feature is a Technology Preview.) Storage ------- * Red Hat Enterprise Linux 6.3 provides full support for Fibre Channel over Ethernet (FCoE) Target. This feature, which was previously provided as a Technology Preview, allows customers to present their Red Hat Enterprise Linux servers as FCoE storage devices. This feature complements the FCoE Initiator support that was delivered in Red Hat Enterprise Linux 6.0. * The Logical Volume Manager (LVM) now provides support for RAID levels 4, 5, and 6. (Previously, support for these RAID levels was provided through the MD subsystem.) This expanded LVM RAID support simplifies overall storage administration by consolidating all management functions, such as creating volumes, resizing volumes, deploying RAID, taking snapshots, etc., into a single interface. (This feature is a Technology Preview.) * The LVM now provides the ability to create thin provisioned logical volumes. Previously, storage was allocated when the volume was created, and needed to be monitored for space consumption and expanded manually. In Red Hat Enterprise Linux 6.3, storage is allocated as required, allowing volumes to expand up to the requested size on demand without intervention. (This feature is a Technology Preview.) Security -------- * Availability of a two-factor authentication mechanism, enhancing the overall security available to lock down Red Hat Enterprise Linux environments and enabling compliance with industry standards such as PCI-DSS. * Expansion of the Advanced Encryption Standard (AES) to provide particular benefits for system performance on multi-processor machines. Identity Management ------------------- * With native support for netgroups and the services map in System Security Services Daemon (SSSD), Red Hat Enterprise Linux servers can be integrated into centralized systems -- such as Active Directory -- to manage system users. * The addition of an automembership plug-in streamlines the administration of new users and hosts when they are added into the Identity Management system by automatically placing them into predefined set of groups, speeding user and host provisioning. * Performance improvements through session data caching, which lowers the overall load on authentication servers. Hardware Enablement ------------------- * Software bandwidth management for USB 3.0 for select Intel platforms is now available. * Compiler optimization for Intel Xeon E5 processor family, which improves the result of string operations, is now included. * Improvements to memory and I/O breakpoint execution operations within compiler tools are now included. Developer Tools --------------- * With the introduction of OpenJDK 7, customers can develop and test with the latest version of open source Java. We look forward to our customer feedback on this beta, and to the continued adoption of Red Hat Enterprise Linux 6, the modern operating platform for the next generation enterprise. To access and download the Beta for Red Hat Enterprise Linux 6.3, please visit: https://access.redhat.com/downloads/ For access to the Red Hat Enterprise Linux 6.3 release notes, please visit: https://docs.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/6-Beta/html/6.3_Release_Notes/index.html For access to all other Red Hat Enterprise Linux 6 documentation, please visit: https://access.redhat.com/knowledge/docs/Red_Hat_Enterprise_Linux/ Sincerely, The Red Hat Enterprise Linux team From michael.coffman at avagotech.com Thu Apr 26 20:54:15 2012 From: michael.coffman at avagotech.com (Michael Coffman) Date: Thu, 26 Apr 2012 14:54:15 -0600 Subject: [rhelv6-list] Kickstart question Message-ID: Any idea why I am seeing the following errors in my anaconda.log file after a kickstart installation: anaconda.log:14:30:48,875 WARNING : '/usr/sbin/hwclock' specified as full path anaconda.log:14:35:34,717 WARNING : '/usr/sbin/authconfig' specified as full path anaconda.log:14:35:35,501 WARNING : '/usr/sbin/lokkit' specified as full path anaconda.log:14:35:35,556 WARNING : '/usr/sbin/lokkit' specified as full path anaconda.log:14:35:36,608 WARNING : '/sbin/grub-install' specified as full path anaconda.log:14:35:36,776 WARNING : '/sbin/grub' specified as full path anaconda.log:14:35:40,483 WARNING : '/bin/sh' specified as full path anaconda.log:14:36:31,285 WARNING : '/bin/sh' specified as full path The /bin/sh happens for each call of %post. I assume the others are similar. I know this doesn't hurt anything, but would still like to know the cause and either fix it or squelch the warnings. I am installing rhel6u2 via PXE Thanks. -- -MichaelC -------------- next part -------------- An HTML attachment was scrubbed... URL: From torriem at gmail.com Sat Apr 28 03:32:09 2012 From: torriem at gmail.com (Michael Torrie) Date: Fri, 27 Apr 2012 21:32:09 -0600 Subject: [rhelv6-list] Red Hat Enterprise Linux 6.3 Beta Announcement In-Reply-To: <4F96CB66.4030409@redhat.com> References: <4F96CB66.4030409@redhat.com> Message-ID: <4F9B64B9.3030203@gmail.com> On 04/24/2012 09:48 AM, Red Hat Enterprise Linux 6 (Santiago) discussion mailing-list wrote: > Virtualization > -------------- > * A new tool called Virt-P2V that facilitates the conversion of physical > Windows or Red Hat Enterprise Linux systems into virtual images to be > deployed as KVM guests inside Red Hat Enterprise Linux or Red Hat > Enterprise Virtualization. > * Stronger compliance with Payment Card Industry Data Security Standards > (PCI-DSS) standards, including the ability to perform secure wipes of > virtual machine disks. > * The ability to perform live volume resizing, improving overall > availability of virtualized guests. Is proper USB 2.0 support in libvirt coming soon? I've got a project that's on hold until this support finds its way into libvirt. I realize this is not specific to RHEL (upstream as it were), but the RHEL release docs said back when 6.0 came out that USB 2.0 support was in the pipeline. From paul at mad-scientist.net Sun Apr 29 22:03:15 2012 From: paul at mad-scientist.net (Paul Smith) Date: Sun, 29 Apr 2012 18:03:15 -0400 Subject: [rhelv6-list] dhclient running TWICE during boot (RHEL 6.2) Message-ID: <1335736995.2044.15.camel@homebase> Hi all; I have a number of up-to-date RHEL 6.2 systems and on all of them I'm seeing a strange issue. I have configured my network interface (using NetworkManager GUI tool) to use DHCP to obtain an IP address. I've also added DHCP_HOSTNAME=myhost to the ifcfg-Auto_eth0 file so that the DNS values are updated for that hostname properly. My DHCP system is configured to handle this properly and indeed it works fine. Here's the problem: during the boot of the system, dhclient is invoked TWICE for the same network interface. Each of these requests obtains its own lease, but for some reason they both get DIFFERENT IP addresses. This happens on all the systems that I've checked (I can see they all have two lease files in /var/lib/dhclient and these lease files have different IP addresses). For some systems it all works out in the end and the interface and the DNS server end up using the same IP address. For other systems, the network interface ends up using the other IP address, not the one that the DNS server is using! Maybe it's a timing thing? I've tricked out one of my systems by renaming the dhclient binary and replacing it with a shell script that generates debugging information and then invokes the real dhclient. I can see that the first invocation of dhclient is performed by rc via S10network. This one uses the default configuration of dhclient which runs the dhclient-script and creates a lease file /var/lib/dhclient/dhclient-eth0.leases. The second invocation is performed by Networkmanager. This one uses a special-crafted configuration of dhclient which runs the nm-dhcp-client.action command and creates a lease file /var/lib/dhclient/dhclient-$UUID-eth0.leases, where $UUID is the value of the UUID variable set in the /etc/sysconfig/network-scripts/ifcfg-Auto_eth0 file. I don't understand why this is happening: if NetworkManager is going to manage the interfaces, why does the network script also do it? I don't know why I get two different IP addresses for the same interface, with the same MAC, either but if I could solve the first problem so that only one dhclient was run, presumably everything would work. I'd prefer to keep the NetworkManager configuration because that's how the users want to manage their network interface (via the graphical tool). Can anyone explain how this is supposed to work and where I've gone wrong? Do other people see duplicated lease files for the same interface when using DHCP? From thomas at redhat.com Mon Apr 30 05:43:10 2012 From: thomas at redhat.com (thomas at redhat.com) Date: Sun, 29 Apr 2012 22:43:10 -0700 Subject: [rhelv6-list] dhclient running TWICE during boot (RHEL 6.2) In-Reply-To: <1335736995.2044.15.camel@homebase> References: <1335736995.2044.15.camel@homebase> Message-ID: <4F9E266E.10702@redhat.com> On 04/29/2012 03:03 PM, Paul Smith wrote: > Hi all; I have a number of up-to-date RHEL 6.2 systems and on all of > them I'm seeing a strange issue. I have configured my network interface > (using NetworkManager GUI tool) to use DHCP to obtain an IP address. > I've also added DHCP_HOSTNAME=myhost to the ifcfg-Auto_eth0 file so that > the DNS values are updated for that hostname properly. > > My DHCP system is configured to handle this properly and indeed it works > fine. > > Here's the problem: during the boot of the system, dhclient is invoked > TWICE for the same network interface. Each of these requests obtains > its own lease, but for some reason they both get DIFFERENT IP addresses. > This happens on all the systems that I've checked (I can see they all > have two lease files in /var/lib/dhclient and these lease files have > different IP addresses). > > For some systems it all works out in the end and the interface and the > DNS server end up using the same IP address. For other systems, the > network interface ends up using the other IP address, not the one that > the DNS server is using! Maybe it's a timing thing? > > I've tricked out one of my systems by renaming the dhclient binary and > replacing it with a shell script that generates debugging information > and then invokes the real dhclient. > > I can see that the first invocation of dhclient is performed by rc via > S10network. This one uses the default configuration of dhclient which > runs the dhclient-script and creates a lease file > /var/lib/dhclient/dhclient-eth0.leases. > > The second invocation is performed by Networkmanager. This one uses a > special-crafted configuration of dhclient which runs the > nm-dhcp-client.action command and creates a lease file > /var/lib/dhclient/dhclient-$UUID-eth0.leases, where $UUID is the value > of the UUID variable set in the > /etc/sysconfig/network-scripts/ifcfg-Auto_eth0 file. > > > I don't understand why this is happening: if NetworkManager is going to > manage the interfaces, why does the network script also do it? I don't > know why I get two different IP addresses for the same interface, with > the same MAC, either but if I could solve the first problem so that only > one dhclient was run, presumably everything would work. I'd prefer to > keep the NetworkManager configuration because that's how the users want > to manage their network interface (via the graphical tool). > > Can anyone explain how this is supposed to work and where I've gone > wrong? Do other people see duplicated lease files for the same > interface when using DHCP? What does /etc/sysconfig/network-scripts/ifcfg-eth0 look like? I generally set ONBOOT=no if I'm going to let NM manage the interface. What happens if you set ONBOOT=no and NM_CONTROLLED=yes? Do you still see the two dhclient processes? -- Thomas Cameron, RHCA, RHCSS, RHCDS, RHCVA, RHCX, CNE, MCSE, MCT Chief Architect, Central US & Canada 512-241-0774 office 512-585-5631 cell 512-857-1345 fax http://people.redhat.com/tcameron For the 5th year running, JBoss leads in customer satisfaction: http://www.jboss.com/pdf/customer_satisfaction.pdf From christian.masopust at siemens.com Mon Apr 30 06:56:48 2012 From: christian.masopust at siemens.com (Masopust, Christian) Date: Mon, 30 Apr 2012 08:56:48 +0200 Subject: [rhelv6-list] Your opinion regarding NFS vs. iSCSI Message-ID: Hi all, I'm going to plan the setup of a database-server (MySQL) and now a discussion started about how the storage should be connected. Some favour iSCSI, others NFS (V4). What's your opinion? Where are advantages / disadvantages? Which solution would promise most performance? Thanks a lot, Christian -------------- next part -------------- An HTML attachment was scrubbed... URL: From paul at mad-scientist.net Mon Apr 30 06:59:12 2012 From: paul at mad-scientist.net (Paul Smith) Date: Mon, 30 Apr 2012 02:59:12 -0400 Subject: [rhelv6-list] dhclient running TWICE during boot (RHEL 6.2) In-Reply-To: <4F9E266E.10702@redhat.com> References: <1335736995.2044.15.camel@homebase> <4F9E266E.10702@redhat.com> Message-ID: <1335769152.11437.6.camel@homebase> On Sun, 2012-04-29 at 22:43 -0700, thomas at redhat.com wrote: > > Can anyone explain how this is supposed to work and where I've gone > > wrong? Do other people see duplicated lease files for the same > > interface when using DHCP? > > What does /etc/sysconfig/network-scripts/ifcfg-eth0 look like? I > generally set ONBOOT=no if I'm going to let NM manage the interface. > What happens if you set ONBOOT=no and NM_CONTROLLED=yes? Do you still > see the two dhclient processes? It's ifcfg-Auto_eth0 actually, since it was created by NM. And it has ONBOOT=yes. It looks like this: HWADDR=05:35:45:D5:F5:35 TYPE=Ethernet BOOTPROTO=dhcp DEFROUTE=yes PEERROUTES=yes IPV4_FAILURE_FATAL=yes IPV6INIT=no NAME="Auto eth3" UUID=75a5eeaf-c5b8-45af-b5e9-e95589be7a35 ONBOOT=yes DHCP_HOSTNAME=myhost I was assuming that if ONBOOT was not set properly, then the interface would not come up at boot time. But maybe NetworkManager doesn't pay attention to that value? If so why is it there (note the above file was created by NetworkManager, not by me). I've never heard of NM_CONTROLLED before; is this documented somewhere? I didn't see it in the docs. Looking through what the scripts do, it doesn't seem to me that this variable has much use. It seems to set USE_NM to true, but that's only used (as far as I can tell) to set the UUID variable if we don't have one... which we do (see above). From jussi_rhel6 at silvennoinen.net Mon Apr 30 07:07:32 2012 From: jussi_rhel6 at silvennoinen.net (Jussi Silvennoinen) Date: Mon, 30 Apr 2012 10:07:32 +0300 (EEST) Subject: [rhelv6-list] Your opinion regarding NFS vs. iSCSI In-Reply-To: References: Message-ID: > Hi all, > ? > I'm going to plan the setup of a database-server (MySQL) and now a > discussion started about > how the storage should be connected. Some favour iSCSI, others NFS (V4). > ? > What's your opinion? Where are advantages / disadvantages? Which solution > would promise > most performance? Curious, SAN is not in your list at all. Why? How important is your service availability to you? -- Jussi From christian.masopust at siemens.com Mon Apr 30 07:22:26 2012 From: christian.masopust at siemens.com (Masopust, Christian) Date: Mon, 30 Apr 2012 09:22:26 +0200 Subject: [rhelv6-list] Your opinion regarding NFS vs. iSCSI In-Reply-To: References: Message-ID: > > Hi all, > > ? > > I'm going to plan the setup of a database-server (MySQL) and now a > > discussion started about > > how the storage should be connected. Some favour iSCSI, > others NFS (V4). > > ? > > What's your opinion? Where are advantages / disadvantages? > Which solution > > would promise > > most performance? > > Curious, SAN is not in your list at all. Why? > How important is your service availability to you? > Hi Jussi, it's also in discussion :) And sure, the service IS important, database will be for mailbox-servers (Zarafa). Currently we're focusing on iSCSI vs. NFS as we don't have FC-equipment but already have 10Gbit ethernet.. Thanks, christian From brilong at cisco.com Mon Apr 30 12:00:36 2012 From: brilong at cisco.com (Brian Long (brilong)) Date: Mon, 30 Apr 2012 12:00:36 +0000 Subject: [rhelv6-list] dhclient running TWICE during boot (RHEL 6.2) In-Reply-To: <1335769152.11437.6.camel@homebase> References: <1335736995.2044.15.camel@homebase> <4F9E266E.10702@redhat.com> <1335769152.11437.6.camel@homebase> Message-ID: Are you sure "chkconfig network off" has been run so NetworkManager owns the network configuration? It almost sounds like both "network" and "NetworkManager" are chkconfig'd on. /Brian/ -- Brian Long | | Corporate Security Programs Org . | | | . | | | . ' ' C I S C O On Apr 30, 2012, at 2:59 AM, Paul Smith wrote: > On Sun, 2012-04-29 at 22:43 -0700, thomas at redhat.com wrote: >>> Can anyone explain how this is supposed to work and where I've gone >>> wrong? Do other people see duplicated lease files for the same >>> interface when using DHCP? >> >> What does /etc/sysconfig/network-scripts/ifcfg-eth0 look like? I >> generally set ONBOOT=no if I'm going to let NM manage the interface. >> What happens if you set ONBOOT=no and NM_CONTROLLED=yes? Do you still >> see the two dhclient processes? > > It's ifcfg-Auto_eth0 actually, since it was created by NM. And it has > ONBOOT=yes. It looks like this: > > HWADDR=05:35:45:D5:F5:35 > TYPE=Ethernet > BOOTPROTO=dhcp > DEFROUTE=yes > PEERROUTES=yes > IPV4_FAILURE_FATAL=yes > IPV6INIT=no > NAME="Auto eth3" > UUID=75a5eeaf-c5b8-45af-b5e9-e95589be7a35 > ONBOOT=yes > DHCP_HOSTNAME=myhost > > I was assuming that if ONBOOT was not set properly, then the interface > would not come up at boot time. But maybe NetworkManager doesn't pay > attention to that value? If so why is it there (note the above file was > created by NetworkManager, not by me). > > I've never heard of NM_CONTROLLED before; is this documented somewhere? > I didn't see it in the docs. Looking through what the scripts do, it > doesn't seem to me that this variable has much use. It seems to set > USE_NM to true, but that's only used (as far as I can tell) to set the > UUID variable if we don't have one... which we do (see above). > > _______________________________________________ > rhelv6-list mailing list > rhelv6-list at redhat.com > https://www.redhat.com/mailman/listinfo/rhelv6-list From o.h.weiergraeber at fz-juelich.de Mon Apr 30 12:44:47 2012 From: o.h.weiergraeber at fz-juelich.de (=?iso-8859-1?Q?=22Weiergr=E4ber=2C_Oliver_H=2E=22?=) Date: Mon, 30 Apr 2012 12:44:47 +0000 Subject: [rhelv6-list] dhclient running TWICE during boot (RHEL 6.2) In-Reply-To: References: <1335736995.2044.15.camel@homebase> <4F9E266E.10702@redhat.com> <1335769152.11437.6.camel@homebase>, Message-ID: Are you sure "chkconfig network off" has been run so NetworkManager owns the network configuration? It almost sounds like both "network" and "NetworkManager" are chkconfig'd on. For reasons I do not understand, this is exactly the default in RHEL 6 - network connections is jointly managed by network and NetworkManager. I have also seen strange behavior under certain conditions which was immediately rectified by turning one of the two services off! IMHO users should be urged to decide which one to use - running both seems to work most of the time, but is prone to errors... Oliver /Brian/ -- Brian Long | | Corporate Security Programs Org . | | | . | | | . ' ' C I S C O On Apr 30, 2012, at 2:59 AM, Paul Smith wrote: > On Sun, 2012-04-29 at 22:43 -0700, thomas at redhat.com wrote: >>> Can anyone explain how this is supposed to work and where I've gone >>> wrong? Do other people see duplicated lease files for the same >>> interface when using DHCP? >> >> What does /etc/sysconfig/network-scripts/ifcfg-eth0 look like? I >> generally set ONBOOT=no if I'm going to let NM manage the interface. >> What happens if you set ONBOOT=no and NM_CONTROLLED=yes? Do you still >> see the two dhclient processes? > > It's ifcfg-Auto_eth0 actually, since it was created by NM. And it has > ONBOOT=yes. It looks like this: > > HWADDR=05:35:45:D5:F5:35 > TYPE=Ethernet > BOOTPROTO=dhcp > DEFROUTE=yes > PEERROUTES=yes > IPV4_FAILURE_FATAL=yes > IPV6INIT=no > NAME="Auto eth3" > UUID=75a5eeaf-c5b8-45af-b5e9-e95589be7a35 > ONBOOT=yes > DHCP_HOSTNAME=myhost > > I was assuming that if ONBOOT was not set properly, then the interface > would not come up at boot time. But maybe NetworkManager doesn't pay > attention to that value? If so why is it there (note the above file was > created by NetworkManager, not by me). > > I've never heard of NM_CONTROLLED before; is this documented somewhere? > I didn't see it in the docs. Looking through what the scripts do, it > doesn't seem to me that this variable has much use. It seems to set > USE_NM to true, but that's only used (as far as I can tell) to set the > UUID variable if we don't have one... which we do (see above). > > _______________________________________________ > rhelv6-list mailing list > rhelv6-list at redhat.com > https://www.redhat.com/mailman/listinfo/rhelv6-list _______________________________________________ rhelv6-list mailing list rhelv6-list at redhat.com https://www.redhat.com/mailman/listinfo/rhelv6-list ------------------------------------------------------------------------------------------------ ------------------------------------------------------------------------------------------------ Forschungszentrum Juelich GmbH 52425 Juelich Sitz der Gesellschaft: Juelich Eingetragen im Handelsregister des Amtsgerichts Dueren Nr. HR B 3498 Vorsitzender des Aufsichtsrats: MinDir Dr. Karl Eugen Huthmacher Geschaeftsfuehrung: Prof. Dr. Achim Bachem (Vorsitzender), Karsten Beneke (stellv. Vorsitzender), Prof. Dr.-Ing. Harald Bolt, Prof. Dr. Sebastian M. Schmidt ------------------------------------------------------------------------------------------------ ------------------------------------------------------------------------------------------------ Kennen Sie schon unsere app? http://www.fz-juelich.de/app From paul at mad-scientist.net Mon Apr 30 12:47:33 2012 From: paul at mad-scientist.net (Paul Smith) Date: Mon, 30 Apr 2012 08:47:33 -0400 Subject: [rhelv6-list] dhclient running TWICE during boot (RHEL 6.2) In-Reply-To: References: <1335736995.2044.15.camel@homebase> <4F9E266E.10702@redhat.com> <1335769152.11437.6.camel@homebase> Message-ID: <1335790053.11437.15.camel@homebase> Thanks for the response! On Mon, 2012-04-30 at 12:00 +0000, Brian Long (brilong) wrote: > Are you sure "chkconfig network off" has been run so NetworkManager > owns the network configuration? It almost sounds like both "network" > and "NetworkManager" are chkconfig'd on. Definitely the "network" service is still enabled. I didn't disable it and apparently nothing else did either. Is that something I need to know to do by hand? I've never heard of that before. It seems very strange to me that you'd have to disable the standard network service in order to use NetworkManager (or disable NM to use the standard network service). What if I wanted some of my interfaces to be managed with NM and others not? From bda20 at cam.ac.uk Mon Apr 30 12:56:09 2012 From: bda20 at cam.ac.uk (Ben) Date: Mon, 30 Apr 2012 13:56:09 +0100 (BST) Subject: [rhelv6-list] dhclient running TWICE during boot (RHEL 6.2) In-Reply-To: <1335790053.11437.15.camel@homebase> References: <1335736995.2044.15.camel@homebase> <4F9E266E.10702@redhat.com> <1335769152.11437.6.camel@homebase> <1335790053.11437.15.camel@homebase> Message-ID: On Mon, 30 Apr 2012, Paul Smith wrote: > It seems very strange to me that you'd have to disable the standard > network service in order to use NetworkManager (or disable NM to use the > standard network service). What if I wanted some of my interfaces to be > managed with NM and others not? Isn't that what NM_CONTROLLED="no" is for in ifcfg-ethN? Ben -- Unix Support, MISD, University of Cambridge, England From wolfy at nobugconsulting.ro Mon Apr 30 12:56:10 2012 From: wolfy at nobugconsulting.ro (Manuel Wolfshant) Date: Mon, 30 Apr 2012 15:56:10 +0300 Subject: [rhelv6-list] dhclient running TWICE during boot (RHEL 6.2) In-Reply-To: References: <1335736995.2044.15.camel@homebase> <4F9E266E.10702@redhat.com> <1335769152.11437.6.camel@homebase>, Message-ID: <4F9E8BEA.5060501@nobugconsulting.ro> On 04/30/2012 03:44 PM, "Weiergr?ber, Oliver H." wrote: > Are you sure "chkconfig network off" has been run so NetworkManager owns the network configuration? It almost sounds like both "network" and "NetworkManager" are chkconfig'd on. > > For reasons I do not understand, this is exactly the default in RHEL 6 - network connections is jointly managed by network and NetworkManager. > I have also seen strange behavior under certain conditions which was immediately rectified by turning one of the two services off! > IMHO users should be urged to decide which one to use - running both seems to work most of the time, but is prone to errors... Both services can coexist but NetworkManager must be told to not handle the interfaces which are managed by the network service ( via the NM_CONTROLLED parameter ) Closest think resembling to docs that I could find is http://docs.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/6-Beta/html/Deployment_Guide/s2-networkscripts-interfaces_network-bridge.html but the parameter is there since Fedora 9 or 10. manuel >> On Sun, 2012-04-29 at 22:43 -0700, thomas at redhat.com wrote: >>>> Can anyone explain how this is supposed to work and where I've gone >>>> wrong? Do other people see duplicated lease files for the same >>>> interface when using DHCP? >>> What does /etc/sysconfig/network-scripts/ifcfg-eth0 look like? I >>> generally set ONBOOT=no if I'm going to let NM manage the interface. >>> What happens if you set ONBOOT=no and NM_CONTROLLED=yes? Do you still >>> see the two dhclient processes? >> It's ifcfg-Auto_eth0 actually, since it was created by NM. And it has >> ONBOOT=yes. It looks like this: >> >> HWADDR=05:35:45:D5:F5:35 >> TYPE=Ethernet >> BOOTPROTO=dhcp >> DEFROUTE=yes >> PEERROUTES=yes >> IPV4_FAILURE_FATAL=yes >> IPV6INIT=no >> NAME="Auto eth3" >> UUID=75a5eeaf-c5b8-45af-b5e9-e95589be7a35 >> ONBOOT=yes >> DHCP_HOSTNAME=myhost >> >> I was assuming that if ONBOOT was not set properly, then the interface >> would not come up at boot time. But maybe NetworkManager doesn't pay >> attention to that value? If so why is it there (note the above file was >> created by NetworkManager, not by me). >> >> I've never heard of NM_CONTROLLED before; is this documented somewhere? >> I didn't see it in the docs. Looking through what the scripts do, it >> doesn't seem to me that this variable has much use. It seems to set >> USE_NM to true, but that's only used (as far as I can tell) to set the >> UUID variable if we don't have one... which we do (see above). >> From paul at mad-scientist.net Mon Apr 30 13:02:02 2012 From: paul at mad-scientist.net (Paul Smith) Date: Mon, 30 Apr 2012 09:02:02 -0400 Subject: [rhelv6-list] dhclient running TWICE during boot (RHEL 6.2) In-Reply-To: References: <1335736995.2044.15.camel@homebase> <4F9E266E.10702@redhat.com> <1335769152.11437.6.camel@homebase> <1335790053.11437.15.camel@homebase> Message-ID: <1335790922.11437.18.camel@homebase> On Mon, 2012-04-30 at 13:56 +0100, Ben wrote: > On Mon, 30 Apr 2012, Paul Smith wrote: > > It seems very strange to me that you'd have to disable the standard > > network service in order to use NetworkManager (or disable NM to use the > > standard network service). What if I wanted some of my interfaces to be > > managed with NM and others not? > > Isn't that what > NM_CONTROLLED="no" > is for in ifcfg-ethN? Well I can't find that flag described in the RHEL documentation, but even so the suggestion was to disable the "network" service. If I disable "network", and I set "NM_CONTROLLED=no", then what will handle that interface? That's what I'm trying to say: I don't think that completely disabling "network" is the right way to solve this problem. From brilong at cisco.com Mon Apr 30 13:07:43 2012 From: brilong at cisco.com (Brian Long (brilong)) Date: Mon, 30 Apr 2012 13:07:43 +0000 Subject: [rhelv6-list] dhclient running TWICE during boot (RHEL 6.2) In-Reply-To: <1335790922.11437.18.camel@homebase> References: <1335736995.2044.15.camel@homebase> <4F9E266E.10702@redhat.com> <1335769152.11437.6.camel@homebase> <1335790053.11437.15.camel@homebase> <1335790922.11437.18.camel@homebase> Message-ID: <4DA0AD92-A702-4125-A8F9-41D03A2936C6@cisco.com> On Apr 30, 2012, at 9:02 AM, Paul Smith wrote: > On Mon, 2012-04-30 at 13:56 +0100, Ben wrote: >> On Mon, 30 Apr 2012, Paul Smith wrote: >>> It seems very strange to me that you'd have to disable the standard >>> network service in order to use NetworkManager (or disable NM to use the >>> standard network service). What if I wanted some of my interfaces to be >>> managed with NM and others not? >> >> Isn't that what >> NM_CONTROLLED="no" >> is for in ifcfg-ethN? > > Well I can't find that flag described in the RHEL documentation, but > even so the suggestion was to disable the "network" service. If I > disable "network", and I set "NM_CONTROLLED=no", then what will handle > that interface? That's what I'm trying to say: I don't think that > completely disabling "network" is the right way to solve this problem. If you want to leave network and NetworkManager enabled on boot, you could set "ONBOOT=no" in your ifcfg-Auth-eth0 file. I believe ONBOOT is only used by the network service and "NM_CONTROLLED" is used by NetworkManager. I'm fairly certain that if ONBOOT=no, NetworkManager will still try to enable the interface depending on the GUI settings. /Brian/ From paul at mad-scientist.net Mon Apr 30 13:15:50 2012 From: paul at mad-scientist.net (Paul Smith) Date: Mon, 30 Apr 2012 09:15:50 -0400 Subject: [rhelv6-list] dhclient running TWICE during boot (RHEL 6.2) In-Reply-To: <4F9E8BEA.5060501@nobugconsulting.ro> References: <1335736995.2044.15.camel@homebase> <4F9E266E.10702@redhat.com> <1335769152.11437.6.camel@homebase> , <4F9E8BEA.5060501@nobugconsulting.ro> Message-ID: <1335791750.11437.29.camel@homebase> On Mon, 2012-04-30 at 15:56 +0300, Manuel Wolfshant wrote: > On 04/30/2012 03:44 PM, "Weiergr?ber, Oliver H." wrote: > > For reasons I do not understand, this is exactly the default in RHEL > > 6 - network connections is jointly managed by network and > > NetworkManager. > > I have also seen strange behavior under certain conditions which was > > immediately rectified by turning one of the two services off! > > IMHO users should be urged to decide which one to use - running both > > seems to work most of the time, but is prone to errors... > Both services can coexist but NetworkManager must be told to not > handle the interfaces which are managed by the network service ( via > the NM_CONTROLLED parameter ) But my problem is really the exact opposite of this. That parameter tells NM to not control an interface that I want managed by the traditional "network" service (apparently...) My problem is the interface is controlled by both and I want it controlled by NM. How do I tell the network service to NOT try to control it? Is setting NM_CONTROLLED=yes sufficient to tell the network service to ignore that interface? If so why doesn't it get set by NM when it starts to manage the interface? I've looked at the scripting for the "network" service and I don't see where setting that parameter will help. On Mon, 30 Apr 2012 13:07:43 +0000, Brian Long wrote: > If you want to leave network and NetworkManager enabled on boot, you > could set "ONBOOT=no" in your ifcfg-Auth-eth0 file. I believe ONBOOT > is only used by the network service and "NM_CONTROLLED" is used by > NetworkManager. I'm fairly certain that if ONBOOT=no, NetworkManager > will still try to enable the interface depending on the GUI settings. OK, I'll try this. I'd really just like to understand what the "correct" method is for this; how Red Hat designed it to work. A little documentation describing these flags and how the two services (SysV network vs. NetworkManager) are intended to interact would be very nice. From notting at redhat.com Mon Apr 30 14:04:24 2012 From: notting at redhat.com (Bill Nottingham) Date: Mon, 30 Apr 2012 10:04:24 -0400 Subject: [rhelv6-list] dhclient running TWICE during boot (RHEL 6.2) In-Reply-To: <1335791750.11437.29.camel@homebase> References: <1335736995.2044.15.camel@homebase> <4F9E266E.10702@redhat.com> <1335769152.11437.6.camel@homebase> <4F9E8BEA.5060501@nobugconsulting.ro> <1335791750.11437.29.camel@homebase> Message-ID: <20120430140424.GB9941@nostromo.devel.redhat.com> Paul Smith (paul at mad-scientist.net) said: > My problem is the interface is controlled by both and I want it > controlled by NM. How do I tell the network service to NOT try to > control it? Is setting NM_CONTROLLED=yes sufficient to tell the network > service to ignore that interface? If so why doesn't it get set by NM > when it starts to manage the interface? I've looked at the scripting > for the "network" service and I don't see where setting that parameter > will help. The network script will use NetworkManager if it's running and redirect ifup to a call to NM to start the interface. However, if NM isn't running yet, it will bring the interface up itself. That being said, NM *should* just take over the valid lease when it starts; if it doesn't, something else is going on. Bill From paul at mad-scientist.net Mon Apr 30 14:29:17 2012 From: paul at mad-scientist.net (Paul Smith) Date: Mon, 30 Apr 2012 10:29:17 -0400 Subject: [rhelv6-list] dhclient running TWICE during boot (RHEL 6.2) In-Reply-To: <20120430140424.GB9941@nostromo.devel.redhat.com> References: <1335736995.2044.15.camel@homebase> <4F9E266E.10702@redhat.com> <1335769152.11437.6.camel@homebase> <4F9E8BEA.5060501@nobugconsulting.ro> <1335791750.11437.29.camel@homebase> <20120430140424.GB9941@nostromo.devel.redhat.com> Message-ID: <1335796157.11437.38.camel@homebase> On Mon, 2012-04-30 at 10:04 -0400, Bill Nottingham wrote: > Paul Smith (paul at mad-scientist.net) said: > > My problem is the interface is controlled by both and I want it > > controlled by NM. How do I tell the network service to NOT try to > > control it? Is setting NM_CONTROLLED=yes sufficient to tell the network > > service to ignore that interface? If so why doesn't it get set by NM > > when it starts to manage the interface? I've looked at the scripting > > for the "network" service and I don't see where setting that parameter > > will help. > > The network script will use NetworkManager if it's running and redirect > ifup to a call to NM to start the interface. However, if NM isn't running > yet, it will bring the interface up itself. During boot, when S10network is run, NM is not running yet... > That being said, NM *should* just take over the valid lease when it starts; > if it doesn't, something else is going on. Not sure what you mean, technically, by "take over the valid lease". Do you mean, when NM starts its dhclient instance it gets back the same results from the DHCP server (since it sent the request with the same MAC), and so both invocations of dhclient have the same result? I do notice only one dhclient is running once the system is up, and ISTR seeing /var/log/messages entries to the effect that once was being killed, so that part is working, but the leases are definitely NOT the same. I don't know why the DHCP server is providing different IP addresses but this is a very widespread phenomena on my network; all the RHEL 6.2 systems I've looked at, even ones where the IP address and DNS entries match, have both these lease files and the IP addresses are different. I believe that our IT department is using some Windows server based DHCP service; maybe it behaves differently than other DHCP servers and if you request a new lease before the previous one was ACK'd it hands out a new IP address, or something. I guess I can try running tcpdump during boot to try to figure out what's going on there. From jussi_rhel6 at silvennoinen.net Mon Apr 30 14:59:11 2012 From: jussi_rhel6 at silvennoinen.net (Jussi Silvennoinen) Date: Mon, 30 Apr 2012 17:59:11 +0300 (EEST) Subject: [rhelv6-list] Your opinion regarding NFS vs. iSCSI In-Reply-To: References: Message-ID: >>> Hi all, >>> ? >>> I'm going to plan the setup of a database-server (MySQL) and now a >>> discussion started about >>> how the storage should be connected. Some favour iSCSI, >> others NFS (V4). >>> ? >>> What's your opinion? Where are advantages / disadvantages? >> Which solution >>> would promise >>> most performance? >> >> Curious, SAN is not in your list at all. Why? >> How important is your service availability to you? > > Hi Jussi, > > it's also in discussion :) And sure, the service IS important, database > will be for mailbox-servers (Zarafa). > > Currently we're focusing on iSCSI vs. NFS as we don't have FC-equipment > but already have 10Gbit ethernet.. I've gotten in to my flame retardant gear so here goes. Ethernet ?s single fabric meaning a single admin error or unexpected end result of plugging new gear to it can bring the whole shebang down. Post-failure less than joyful consistency check marathon is sure to follow. For me, that is unacceptable. I'd rather be enjoying my beer at the local pub instead. FC SAN being multi-fabric, you have to try really hard to break everything. Whatever the transport technology is based on, ethernet, FC or snails on steroids, if it has multiple independent fabrics, I'm willing to listen. Otherwise, I'll pass. I really don't see any need or use for FCoE. I do like the idea of a single communications channel (redundant) but FCoE is a poor excuse for a solution towards that. iSCSI is much simpler protocol but suffers the same single fabric shortcoming. Perhaps there are ways out there to do ethernet-based blockstorage reliably that other list members know about, I'd certainly want to know about them. -- Jussi From paul at mad-scientist.net Mon Apr 30 17:47:40 2012 From: paul at mad-scientist.net (Paul Smith) Date: Mon, 30 Apr 2012 13:47:40 -0400 Subject: [rhelv6-list] dhclient running TWICE during boot (RHEL 6.2) In-Reply-To: <4DA0AD92-A702-4125-A8F9-41D03A2936C6@cisco.com> References: <1335736995.2044.15.camel@homebase> <4F9E266E.10702@redhat.com> <1335769152.11437.6.camel@homebase> <1335790053.11437.15.camel@homebase> <1335790922.11437.18.camel@homebase> <4DA0AD92-A702-4125-A8F9-41D03A2936C6@cisco.com> Message-ID: <1335808060.3891.10.camel@psmith-ubeta> On Mon, 2012-04-30 at 13:07 +0000, Brian Long (brilong) wrote: > If you want to leave network and NetworkManager enabled on boot, you > could set "ONBOOT=no" in your ifcfg-Auth-eth0 file. I believe ONBOOT > is only used by the network service and "NM_CONTROLLED" is used by > NetworkManager. I'm fairly certain that if ONBOOT=no, NetworkManager > will still try to enable the interface depending on the GUI settings. Man this is ... weak. So I tried this. If I set ONBOOT=no and NM_CONTROLLED=yes, then the system boots BUT it never brings up the network interface at all, until I log in at the console. That's not acceptable, I want the network interface to come up automatically at boot so I can access it remotely if it reboots without having to go log into it. So I switched them and set ONBOOT=yes and NM_CONTROLLED=no. When I rebooted this time the network interface came up (with the "wrong" IP address), but no default route was created! So I couldn't access any other hosts. Of course I can do this by hand but again, I don't want to have to to fix the host whenever it reboots. So I set ONBOOT=yes again and removed the NM_CONTROLLED variable and it went back to the old behavior... still broken but at least I can access the system by IP address! Here's even more awesome: looking through the scripts I saw a USE_NM variable, so I tried adding "USE_NM=yes" to my ifcfg-Auto_eth0 script. When I rebooted this time the boot process HUNG when S10network was trying to bring up eth0. I couldn't ^C or ^Z or switch to an alternate console (maybe it's too early for them?). The only keys that would do anything were CTRL-ALT-DEL and that just brought me right back to the same state. Trying to boot into safe mode gave me a kernel panic. Eventually I had to boot off of a separate DVD, mount the partition, remove that line, then I could boot again. Fun! From lowen at pari.edu Mon Apr 30 17:47:24 2012 From: lowen at pari.edu (Lamar Owen) Date: Mon, 30 Apr 2012 13:47:24 -0400 Subject: [rhelv6-list] Your opinion regarding NFS vs. iSCSI In-Reply-To: References: Message-ID: <201204301347.25060.lowen@pari.edu> On Monday, April 30, 2012 10:59:11 AM Jussi Silvennoinen wrote: > I really don't see any need or use for FCoE. I do like the idea of a > single communications channel (redundant) but FCoE is a poor excuse for a > solution towards that. iSCSI is much simpler protocol but suffers the same > single fabric shortcoming. Ethernet is just as easy to deploy multifabric as FC is; you just design with multiple physical switches or sets of switches (just like FC). I wouldn't deploy Ethernet for the SAN on the same physical switches and topology as my LAN anyway, assuming existing switch availability or new switch purchase budget for that task. If the budget is tight, at least try to implement the backup or secondary SAN fabric on its own switches (I'd not put the primary on a dedicated fabric with the backup overlaying the LAN; the backup should always be more reliable than the primary). If using full active/active fabrics, one on its own hardware and links and one overlaying the LAN might be serviceable. Having said that, I am FC multifabric here. But the same overall SAN architecture rules apply with Ethernet as with FC, really, for either iSCSI or FCoE. You can even carry LAN traffic on FC (IPoFC). But to do it because 'we already have 10GbE' seems to indicate that SAN overlay on the LAN is being contemplated, and this is where things can get hairy. You really do want completely separate networks for SAN and LAN, if only to eliminate spanning-tree issues on the LAN creating issues for the SAN. FC has other advantages over the Ethernet technologies, too, but it does cost quite a bit more. IMO, YMMV, etc. From notting at redhat.com Mon Apr 30 18:01:04 2012 From: notting at redhat.com (Bill Nottingham) Date: Mon, 30 Apr 2012 14:01:04 -0400 Subject: [rhelv6-list] dhclient running TWICE during boot (RHEL 6.2) In-Reply-To: <1335808060.3891.10.camel@psmith-ubeta> References: <1335736995.2044.15.camel@homebase> <4F9E266E.10702@redhat.com> <1335769152.11437.6.camel@homebase> <1335790053.11437.15.camel@homebase> <1335790922.11437.18.camel@homebase> <4DA0AD92-A702-4125-A8F9-41D03A2936C6@cisco.com> <1335808060.3891.10.camel@psmith-ubeta> Message-ID: <20120430180104.GA13589@nostromo.devel.redhat.com> Paul Smith (paul at mad-scientist.net) said: > So I tried this. If I set ONBOOT=no and NM_CONTROLLED=yes, then the > system boots BUT it never brings up the network interface at all, until > I log in at the console. That's not acceptable, I want the network > interface to come up automatically at boot so I can access it remotely > if it reboots without having to go log into it. NM honors ONBOOT=no, so this is expected. > So I switched them and set ONBOOT=yes and NM_CONTROLLED=no. When I > rebooted this time the network interface came up (with the "wrong" IP > address), but no default route was created! So I couldn't access any > other hosts. Of course I can do this by hand but again, I don't want to > have to to fix the host whenever it reboots. This means that it's only brought up by /etc/init.d/network; route should be grabbed from the DHCP server. > Here's even more awesome: looking through the scripts I saw a USE_NM > variable, so I tried adding "USE_NM=yes" to my ifcfg-Auto_eth0 script. > When I rebooted this time the boot process HUNG when S10network was > trying to bring up eth0. I couldn't ^C or ^Z or switch to an alternate > console (maybe it's too early for them?). The only keys that would do > anything were CTRL-ALT-DEL and that just brought me right back to the > same state. Trying to boot into safe mode gave me a kernel panic. USE_NM is internal initscripts accounting - I suppose it should be namespaced off. It's not ment as an end-user configuration variable. Bill From geslinux at gmail.com Mon Apr 30 18:14:55 2012 From: geslinux at gmail.com (Grzegorz Witkowski) Date: Mon, 30 Apr 2012 19:14:55 +0100 Subject: [rhelv6-list] Your opinion regarding NFS vs. iSCSI In-Reply-To: References: Message-ID: It is easy and simple to build fully redundant iscsi network which will deliver and cost much less than FC. Also troubleshooting is pretty easy. iSCSI can be a really good choice if the design is right. There are many factors involved. You cannot simply ask "iscsi or fc?" On Apr 30, 2012 4:01 p.m., "Jussi Silvennoinen" < jussi_rhel6 at silvennoinen.net> wrote: > Hi all, >>>> >>>> I'm going to plan the setup of a database-server (MySQL) and now a >>>> discussion started about >>>> how the storage should be connected. Some favour iSCSI, >>>> >>> others NFS (V4). >>> >>>> >>>> What's your opinion? Where are advantages / disadvantages? >>>> >>> Which solution >>> >>>> would promise >>>> most performance? >>>> >>> >>> Curious, SAN is not in your list at all. Why? >>> How important is your service availability to you? >>> >> >> Hi Jussi, >> >> it's also in discussion :) And sure, the service IS important, database >> will be for mailbox-servers (Zarafa). >> >> Currently we're focusing on iSCSI vs. NFS as we don't have FC-equipment >> but already have 10Gbit ethernet.. >> > > I've gotten in to my flame retardant gear so here goes. > > Ethernet ?s single fabric meaning a single admin error or unexpected end > result of plugging new gear to it can bring the whole shebang down. > Post-failure less than joyful consistency check marathon is sure to follow. > > For me, that is unacceptable. I'd rather be enjoying my beer at the local > pub instead. FC SAN being multi-fabric, you have to try really hard to > break everything. > > Whatever the transport technology is based on, ethernet, FC or snails on > steroids, if it has multiple independent fabrics, I'm willing to listen. > Otherwise, I'll pass. > > I really don't see any need or use for FCoE. I do like the idea of a > single communications channel (redundant) but FCoE is a poor excuse for a > solution towards that. iSCSI is much simpler protocol but suffers the same > single fabric shortcoming. > > Perhaps there are ways out there to do ethernet-based blockstorage > reliably that other list members know about, I'd certainly want to know > about them. > > > > -- > > Jussi > _______________________________________________ > rhelv6-list mailing list > rhelv6-list at redhat.com > https://www.redhat.com/mailman/listinfo/rhelv6-list > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From notting at redhat.com Mon Apr 30 19:14:57 2012 From: notting at redhat.com (Bill Nottingham) Date: Mon, 30 Apr 2012 15:14:57 -0400 Subject: [rhelv6-list] dhclient running TWICE during boot (RHEL 6.2) In-Reply-To: <20120430180104.GA13589@nostromo.devel.redhat.com> References: <1335736995.2044.15.camel@homebase> <4F9E266E.10702@redhat.com> <1335769152.11437.6.camel@homebase> <1335790053.11437.15.camel@homebase> <1335790922.11437.18.camel@homebase> <4DA0AD92-A702-4125-A8F9-41D03A2936C6@cisco.com> <1335808060.3891.10.camel@psmith-ubeta> <20120430180104.GA13589@nostromo.devel.redhat.com> Message-ID: <20120430191457.GA14432@nostromo.devel.redhat.com> Bill Nottingham (notting at redhat.com) said: > Paul Smith (paul at mad-scientist.net) said: > > So I tried this. If I set ONBOOT=no and NM_CONTROLLED=yes, then the > > system boots BUT it never brings up the network interface at all, until > > I log in at the console. That's not acceptable, I want the network > > interface to come up automatically at boot so I can access it remotely > > if it reboots without having to go log into it. > > NM honors ONBOOT=no, so this is expected. > > > So I switched them and set ONBOOT=yes and NM_CONTROLLED=no. When I > > rebooted this time the network interface came up (with the "wrong" IP > > address), but no default route was created! So I couldn't access any > > other hosts. Of course I can do this by hand but again, I don't want to > > have to to fix the host whenever it reboots. > > This means that it's only brought up by /etc/init.d/network; route should > be grabbed from the DHCP server. > > > Here's even more awesome: looking through the scripts I saw a USE_NM > > variable, so I tried adding "USE_NM=yes" to my ifcfg-Auto_eth0 script. > > When I rebooted this time the boot process HUNG when S10network was > > trying to bring up eth0. I couldn't ^C or ^Z or switch to an alternate > > console (maybe it's too early for them?). The only keys that would do > > anything were CTRL-ALT-DEL and that just brought me right back to the > > same state. Trying to boot into safe mode gave me a kernel panic. > > USE_NM is internal initscripts accounting - I suppose it should be > namespaced off. It's not ment as an end-user configuration variable. So, I suspect that a lot of this is due to lease file confusion between the initscripts and NetworkManager; they're using different lease file names, so dhclient does DISCOVER instead of just renewing the existing lease. As a workaround in initscripts, you can massage something like: http://git.fedorahosted.org/git/?p=initscripts.git;a=commitdiff;h=60b84547c03ec7d82633f3c9d6939af4747a6bff to apply on RHEL 6, and remove your lease files and reboot. I've filed a bug to have NM adjust to the existing lease file names on RHEL 6: https://bugzilla.redhat.com/show_bug.cgi?id=817660 which would help as well. Bill