On Mon, Feb 16, 2015 at 03:25:58PM +0100, Kristoffer Grönlund wrote: > Dejan Muhamedagic <deja...@fastmail.fm> writes: > > >> > >> If Dejan also votes to revert this change, I will do so. > > > > Must say that I somehow missed the change. The most useful thing > > to do seems be to exit with 0 in case --force (or crm -F) is in > > effect (perhaps to suppress the error message too), just like "rm > > -f", which Lars brought up. Otherwise to keep the existing > > behaviour (error code, error message). > > Fixed. :) > > > > >> As a sidenote, I find the automatic removal of related constraints > >> somewhat strange... > > > > What good could there be from a constraint if the referenced > > resource left the stage? I may have found the idea strange too, > > but then ended up with an illegal CIB, which was not the best way > > forward either. To the best of my knowledge, crm_verify doesn't > > allow constraints pointing to nowhere. > > > > It really is only a problem of expected behavior, especially when > editing. I might choose to modify a primitive by removing and recreating > it, expecting the constraints to still be there. Or I want to remove a > primitive and rewrite constraints applied to it to apply to a different > resource instead, and be surprised when those constraints > disappear.
Yes, I see your point and it could just as well be that crmsh is really overzealous in this case. Perhaps the removal should be deferred until the commit? Or is it that with certain strictness setting CIB verification is run after every configuration change (i.e. before commit)? Cheers, Dejan > But, the current functionality makes sense in light of how > the CIB works. > > -- > // Kristoffer Grönlund > // kgronl...@suse.com _______________________________________________ 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
