Hi all, I was informed[1] that gfs & rhcs packages are in the works but that there are some problems with the code. What are these problems? Can we do anything to help test / solve them? [1] http://www.karan.org/blog/index.php/2005/07/25/everyone_loves_a_cluster_as_do_i -- Jure Pe?ar http://jure.pecar.org
On Mon, 2005-11-21 at 15:31 +0100, Jure Pe?ar wrote:> Hi all, > > I was informed[1] that gfs & rhcs packages are in the works but that there are some problems with the code. What are these problems? Can we do anything to help test / solve them? > > > > [1] http://www.karan.org/blog/index.php/2005/07/25/everyone_loves_a_cluster_as_do_i >We just built some more i686 packages that Karanbir is going to test in the next day or so. If they work even fairly well, we will make them available for testing. -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 189 bytes Desc: This is a digitally signed message part URL: <http://lists.centos.org/pipermail/centos/attachments/20051121/28f07026/attachment-0005.sig>
On Mon, 21 Nov 2005 08:41:19 -0600 Johnny Hughes <mailing-lists at hughesjr.com> wrote:> We just built some more i686 packages that Karanbir is going to test in > the next day or so. > > If they work even fairly well, we will make them available for testing.I'm able to build & test x86_64 stuff, since I'm setting up a small cluster real-soon-now :) -- Jure Pe?ar http://jure.pecar.org
On Mon, 21 Nov 2005 08:41:19 -0600 Johnny Hughes <mailing-lists at hughesjr.com> wrote:> We just built some more i686 packages that Karanbir is going to test in > the next day or so. > > If they work even fairly well, we will make them available for testing.Any progress on this? -- Jure Pe?ar http://jure.pecar.org
From: Karanbir Singh [mailto:mail-lists at karan.org]> > Jure Pecar wrote: > > On Mon, 21 Nov 2005 08:41:19 -0600 > > Johnny Hughes <mailing-lists at hughesjr.com> wrote: > > > >>We just built some more i686 packages that Karanbir is going to test in > >>the next day or so. > >> > >>If they work even fairly well, we will make them available for testing. > > > > Any progress on this? > >http://lists.centos.org/pipermail/centos-announce/2005-December/000946.html I don't remember seeing this. Is the announce list mirrored to this one, or do I need to subscribe to yet another mailing list to get these announcements? Bowie
From: Karanbir Singh [mailto:mail-lists at karan.org]> > Bowie Bailey wrote: > > From: Karanbir Singh [mailto:mail-lists at karan.org] > > > >>Jure Pecar wrote: > >> > >>>On Mon, 21 Nov 2005 08:41:19 -0600 > >>>Johnny Hughes <mailing-lists at hughesjr.com> wrote: > >>> > >>> > >>>>We just built some more i686 packages that Karanbir is going to testin> >>>>the next day or so. > >>>> > >>>>If they work even fairly well, we will make them available fortesting.> >>> > >>>Any progress on this? > >> > >>http://lists.centos.org/pipermail/centos-announce/2005-December/000946.html> > > > I don't remember seeing this. Is the announce list mirrored to this > > one, or do I need to subscribe to yet another mailing list to get > > these announcements? > > the annouce list is posted in digest form to this list, one post per > day , on days when there is a post in the announce list.Ok, I guess I just missed this one. On a question more related to the original subject... I have a server using cs and gfs that is built from the source rpms on http://rpm.karan.org/el4/csgfs/SRPMS/ (modified only enough to compile on the latest kernel without hugemem). Is it safe (or even possible) to update these from the official distribution, or do I need to remove the current stuff and rebuild the cluster using the official rpms? I would like to get to a point where this can be kept up to date with a 'yum update' command now that it's in the CentOS distribution. Bowie
From: Karanbir Singh [mailto:mail-lists at karan.org]> > Bowie Bailey wrote: > > > > I have a server using cs and gfs that is built from the source rpms on > > http://rpm.karan.org/el4/csgfs/SRPMS/ (modified only enough to compile > > on the latest kernel without hugemem). Is it safe (or even possible) > > to update these from the official distribution, or do I need to remove > > the current stuff and rebuild the cluster using the official rpms? > > > > I would like to get to a point where this can be kept up to date with > > a 'yum update' command now that it's in the CentOS distribution. > > You should be able to just setup a repo for the csgfs release on > mirror.centos.org and yum update against that with no problems.I did that and immediately ran into problems. # depmod -a # modprobe dm-mod # modprobe gfs FATAL: Module gfs not found. # modprobe lock_dlm FATAL: Module lock_dlm not found. Everything is installed, but it looks like some of the modules didn't upgrade properly. ccs-1.0.2-0 cman-1.0.2-0 cman-kernel-2.6.9-36.0 cman-kernel-2.6.9-39.8.centos4 cman-kernel-smp-2.6.9-36.0 cman-kernel-smp-2.6.9-39.8.centos4 cman-kernheaders-2.6.9-39.8.centos4 dlm-1.0.0-5 dlm-devel-1.0.0-5 dlm-kernel-2.6.9-34.0 dlm-kernel-2.6.9-37.9.centos4 dlm-kernel-smp-2.6.9-34.0 dlm-kernel-smp-2.6.9-37.9.centos4 dlm-kernheaders-2.6.9-37.9.centos4 fence-1.32.6-0 GFS-6.1.2-0 GFS-kernel-2.6.9-35.5 GFS-kernel-2.6.9-42.2.centos4 GFS-kernel-smp-2.6.9-35.5 GFS-kernel-smp-2.6.9-42.2.centos4 gulm-1.0.4-0 gulm-devel-1.0.4-0 iddev-2.0.0-3 lvm2-cluster-2.01.14-1.0.RHEL4 magma-1.0.1-4 magma-plugins-1.0.2-0 system-config-cluster-1.0.16-1.0 There appear to be two versions of several of the packages. What is the best way to recover from this? Should I remove them all and start over, or is there a better way? Bowie
From: Karanbir Singh [mailto:mail-lists at karan.org]> > Bowie Bailey wrote: > > > > Everything is installed, but it looks like some of the modules > > didn't upgrade properly. > > upgrade ? where are you upgrading from ? This is the first csgfs > release on CentOS4 :) > > > > > There appear to be two versions of several of the packages. > > > > how did you install this in the first place ?I originally installed from the source rpms on your website... You told me that these should update properly from the CentOS distribution using 'yum update', so I tried it. From: Karanbir Singh [mailto:mail-lists at karan.org]> > Bowie Bailey wrote: > > > > I have a server using cs and gfs that is built from the source rpms on > > http://rpm.karan.org/el4/csgfs/SRPMS/ (modified only enough to compile > > on the latest kernel without hugemem). Is it safe (or even possible) > > to update these from the official distribution, or do I need to remove > > the current stuff and rebuild the cluster using the official rpms? > > > > I would like to get to a point where this can be kept up to date with > > a 'yum update' command now that it's in the CentOS distribution. > > You should be able to just setup a repo for the csgfs release on > mirror.centos.org and yum update against that with no problems.In any case, this question is moot now as I have rebuilt from the official packages. I am still having some problems getting things operational, but I will post another question on that later if I can't figure it out. Bowie