On Thu, Oct 11, 2012 at 11:43:40AM +1100, Andrew Beekhof wrote: > On Tue, Sep 25, 2012 at 10:44 PM, Ulrich Windl > <[email protected]> wrote: > > Hi! > > > > I have a question on pacemaker: Can I have a ordering and location > > constraint between a primitive and an instance of a clone instance? I'm > > usure whether or how it works. > > > > Background: If you have a cluster filesystem that is running on n nodes > > (maybe not all of your notes), and you want to run a virtual machine where > > a CFS is running and after the CFS is started. Likewise, if an instance of > > the clones CFS goes down, the dependent VM on that node should also go down. > > > > If my CFS is "cln_cfs_VMs", and my VMs are like "prm_xen_s*", I tried: > > colocation col_Xen_VMs_cfs inf: ( prm_xen_s01 prm_xen_s02 prm_xen_s03 > > prm_xen_s04 prm_xen_s05 prm_xen_s06 prm_xen_s07 ) cln_cfs_VMs > > order ord_cfs_VMs_Xen inf: cln_cfs_VMs ( prm_xen_s01 prm_xen_s02 > > prm_xen_s03 prm_xen_s04 prm_xen_s05 prm_xen_s06 prm_xen_s07 ) > > I've gives up trying to understand the shell syntax for sets... can > you show me the xml?
The shell just reflects what's in the CIB. There's nothing funky in the syntax. > What you describe should be possible, I'm just not sure if you have > the right syntax. FWIW, the colocation above looks OK. Thanks, Dejan > > The CFS is OCFS with ist own constraints: > > order ord_DLM_O2CB inf: cln_DLM cln_O2CB > > order ord_OCFS_fs inf: cln_O2CB ( cln_cfs_locks1 cln_cfs_VMs ) > > > > Is this too complex for CRM to handle, or did I simply make a mistake? > > > > Regards, > > Ulrich > > > > > > _______________________________________________ > > Linux-HA mailing list > > [email protected] > > http://lists.linux-ha.org/mailman/listinfo/linux-ha > > See also: http://linux-ha.org/ReportingProblems > _______________________________________________ > Linux-HA mailing list > [email protected] > http://lists.linux-ha.org/mailman/listinfo/linux-ha > See also: http://linux-ha.org/ReportingProblems _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
