Hi, On Mon, Dec 10, 2007 at 04:10:20PM -0800, Radu Handorean wrote: > Hello all, > > When I force a failover from one node to the other (I need all > my resources on the same node at all times) it might take a > while for all resources to migrate. One of the resources to > migrate is the cluster IP but just because I can connect to the > other node (this resource has migrated) doesn't mean all of > them have. > > > I need a way to figure when the migration is complete. I could > use crm_mon and sleep/loop until all of them are on the other > side (I know how many and what to be looking for). For no good > reason, I feel this is not very elegant. I think I've heard > about resource dependencies and how one can be declared to be > dependent on other(s). If this is true, can someone please > point me to where I can read about it? I'm thinking about > defining a resource to be dependent on all the others and > verifying when this "phony" resource migrated which would > indicate they are all on the other side. Is this a good > approach? Can you give me any other idea for how to decide if > the migration is over or not?
You can test (monitor) the last resource in a group. It is possible to run a resource agent by hand: just set the OCF_RESKEY_<param> environment variables to the corresponding values from your CIB. You can also try crmadmin -S DC_NODE in a loop until it reports the IDLE state, then check whichever resource you want. Thanks, Dejan > Thanks! > > > > > ____________________________________________________________________________________ > Looking for last minute shopping deals? > Find them fast with Yahoo! Search. > http://tools.search.yahoo.com/newsearch/category.php?category=shopping > _______________________________________________ > 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
