On 2011-01-08 16:21, Holger Teutsch wrote: > Hi, > I came across the exact semantics of the "unique" attribute for > parameters. > > Florian's (really great) guide states that > > Finally, the unique attribute (allowed values: true or false) > indicates that a specific value must be unique across the > cluster, for this parameter of this particular resource type > > Experimenting shows that obviously nothings prevents entering a 'unique' > parameter several times in the same cluster.
That is correct. The shell does not enforce uniqueness. > In Andrew's "Pacemaker Explained 1.1" Chap 9.5 the 'unique' attribute is > only tied to an optional 'reload' action. It marks this particular > parameter as reloadable. > > While cursory browsing through the pacemaker source I could not find any > other usage for this attribute. > > Might be the RA guide should state that the name of this attribute is > somewhat misleading and refer to/quote the pacemaker guide. AFAICT the RA dev guide is actually "correct" in the sense that it documents what the OCF spec expects from a resource agent. Note that the dev guide also does not make any statements as to how _any_ OCF RA semantics are enforced at the cluster manager level. Dejan, is there any chance that you could put parameter uniqueness enforcement into the shell? Or maybe this is something that needs to be put in the PE itself, and should flag a warning in ptest output? Cheers, Florian
signature.asc
Description: OpenPGP digital signature
_______________________________________________________ Linux-HA-Dev: [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev Home Page: http://linux-ha.org/
