>>> Andrew Beekhof <and...@beekhof.net> schrieb am 12.08.2011 um 02:29 in >>> Nachricht <CAEDLWG1+To1ED5gZMUj+qPsYsHy7ecBWkgBTEQ=O==a5hze...@mail.gmail.com>: > On Thu, Aug 11, 2011 at 5:08 PM, Ulrich Windl > <ulrich.wi...@rz.uni-regensburg.de> wrote: >>>>> Andrew Beekhof <and...@beekhof.net> schrieb am 11.08.2011 um 07:57 in > Nachricht > > <CAEDLWG3UfkJsYf3x9CUu45K9vdO1rce7FF9V1sooHkdp_X=x...@mail.gmail.com>: > >> On Sat, Aug 6, 2011 at 12:01 AM, Ulrich Windl > >> <ulrich.wi...@rz.uni-regensburg.de> wrote: > >> > Hi! > >> > > >> > I frequently see problems I don't understand: > >> > When configuring an exportfs resource using "crm" shell without > >> > explicitly > >> specifying operations or timeouts, I get warnings like these: > >> > WARNING: prm_nfs_v03: default timeout 20s for start is smaller than the > >> advised 40 > >> > > >> > I wonder: If the default is "40s", > >> > >> It is not the default. > >> It is the recommended minimum for that operation on that resource. > > > > OK, > > > > let's rephrase it: If there is an advertised minimum, and I do not specify > a timeout, why isn't hat advertised minimum used (as a default)? > > Because its only a recommendation > Because you may have configured it outside of the shell - with a tools > which doesn't know about the agent's metadata. > Because the metadata might be different on other machines in the cluster.
Hi! Sorry, but I get more new questions than answers: The operation defaults an in the CIB, while the minimums are in the RA's metadata, right? So any configuration tool should have access to both data. Why shouldn't a tool have access to the RA's metadata? Also why should those metadata be different on individual nodes of one cluster? I'd consider it a bug if different nodes have different versions of the RA installed. Regards, Ulrich _______________________________________________ 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
