Hi, On Thu, Feb 18, 2010 at 05:33:53PM +0100, Cristian Mammoli - Apra Sistemi wrote: > Dejan Muhamedagic wrote: > > >>Last packages from clusterlabs repo: should be ok? > > > >I'm really not sure. See the bugzilla above. The problem is that > >at the time of 1.0.7 the CRM and the shell were out of sync on > >the matter. > > My package has build date: Wed 10 Feb 2010
OK, that should be fine. Just don't use unreleased versions for production. > It should be fixed, indeed the xml syntax of a test resource has the > score only at "rsc_colocation" level. > > >>Where can I find some examples of resource sets through the cli? > > > >The doc on crm cli: http://www.clusterlabs.org/doc/crm_cli.html > >but the resource sets are not described. I'll fix that tomorrow. > > > Thank you > > >>Basically what I want is > >> > >>Promote resource A master on some node > >>Then Start resource B on the same node > >>Then Start resources C,D,E unordered on the same node > >> > >>Is colocation sufficient or i need both an order constraint and a > >>colo constraint? > > > >What you described above is order of operations: > > > >order o1 inf: A:promote B:start ( C D E ) > > http://www.clusterlabs.org/doc/en-US/Pacemaker/1.0/html/Pacemaker_Explained/s-resource-sets-collocation.html > What's the purpose of the last example at chapther 6.6 of "pacemaker > explained" then? Sorry, didn't read careful enough, was distracted by the actions in the description. Yes, you need both the order and the collocation. > Isn't it supposed to do what I want? I guess so. > Basically what I want to achieve is the same behaviour of an > unordered group of heartbeat 2.1.4. Why don't you keep the unordered group? > Isn't there a way to obtain the same? Now if i have 10 unordered > resources I should write 10 colocation constraints and 10 order > constraints instead of 1. I'm a bit confused. You already used an unordered group and resource sets. Shouldn't they do? Thanks, Dejan > >It should look like that. Anyway, you'll probably have to wait for > >1.0.8. Or grab current sources. > > I'll wait for the RPMs (and the documentation as well). > > Bye > > -- > Cristian Mammoli > APRA SISTEMI srl > > > _______________________________________________ > Pacemaker mailing list > Pacemaker@oss.clusterlabs.org > http://oss.clusterlabs.org/mailman/listinfo/pacemaker _______________________________________________ Pacemaker mailing list Pacemaker@oss.clusterlabs.org http://oss.clusterlabs.org/mailman/listinfo/pacemaker
