On 26/10/11 16:42, alain.mou...@bull.net wrote: > Hi Nick, > > I spent e few weeks to try to make this stack ocfs2.pcmk working with > pacemaker/corosync stack > last fall 2010, and this stack was definitely not working on RH. Many > conflicts > between ocfs2 clustering and pacemaker clustering leading to both nodes > killed > or at least ocfs2 fs read-only whereas the other node was fenced, etc. I > give up > on this stack. > I 'm very curious to know if it works now (at least on RHEL6) but I don't > think so > as the response of Oracle last year was that they did not work on this > stack again. > But perhaps they have now ...
You can't (or at least, seriously shouldn't - that way lies madness) have both Pacemaker and O2CB simultaneously managing the cluster. I can't speak for OCFS2 support on Debian or RHEL, but I can suggest reading: * The previous version of Clusters from Scratch (labelled "OpenAIS") at http://www.clusterlabs.org/wiki/Documentation (this is old, but refers to OCFS2). * The relevant chapter of the SLE HA documentation: http://www.novell.com/documentation/sle_ha/book_sleha/?page=/documentation/sle_ha/book_sleha/data/part_storage.html * Lars' blog post about using clone-of-group configuration in Pacemaker: http://www.advogato.org/person/lmb/diary/104.html HTH, Tim > > Let us know. > Thanks a lot > Alain Moullé > > > > De : Nick Khamis<sym...@gmail.com> > A : General Linux-HA mailing list<linux-ha@lists.linux-ha.org> > Date : 25/10/2011 20:14 > Objet : [Linux-HA] PCMK + OCFS2 > Envoyé par : linux-ha-boun...@lists.linux-ha.org > > > > Hello Everyone, > > Moving forward, I noticed that there was not much documentation > regarding getting the pcmk stack working > with ocfs2. I have the configuration up and running however, missed > the part regarding getting what is required > for pcmk+ocfs support (to get ocf:pacemaker:controld + > ocf:pacemaker:o2cb working). > > Everything is build from source using the latest version of Glue, RA, > PCMK, and OpenAIS. OCFS2 works fine > manually, and now I am trying to get corosync to handle it. This is on > a prototype environment right now using > Debian Squeeze however, will be using an EL like Red Hat for > production. No stonith is required just yet, but if > the documentation includes that as well it would be beneficial very soon. > > I did find: > http://oss.clusterlabs.org/pipermail/pacemaker/attachments/20091013/b7980710/attachment.bat > > from > 2009 but reckon a lot has changed since then? > > Thanks in Advance as Always, > > Nick from Toronto. > _______________________________________________ > Linux-HA mailing list > Linux-HA@lists.linux-ha.org > http://lists.linux-ha.org/mailman/listinfo/linux-ha > See also: http://linux-ha.org/ReportingProblems > > _______________________________________________ > Linux-HA mailing list > Linux-HA@lists.linux-ha.org > http://lists.linux-ha.org/mailman/listinfo/linux-ha > See also: http://linux-ha.org/ReportingProblems > _______________________________________________ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems