On Tue Sep 01, 2009 at 12:03:47 +0200, markus prothmann wrote: > >Hi all, > >i am working on a two node squid cluster with heartbeat2 and squid3. >The cluster was installed on debian 5.02 i386. the system is up and >running - what is going well by now is: The two nodes are working as a >cluster with a secondary ip on their lan interface. squid3 is started >from the heartbeat daemon. if i shutdown one node, the squid and the >ip is brought up in the right way on the second node. hard shutdown of >the activ node also moves the resources. I used the main parts from a >prior post (but i attached the squid script and the xml what i >inserted to the debian default cib. > >What is not working correct (imo) is: if you rename the squid.conf on >the active node (so squid is not able to start on it) heartbeat tries >to restart squid but does not move the resources. And the inserted >ping nodes seem to have no effect - disbaling the lan interface of a >active node, does not force the resources to move.
It sounds like you need to set a reasonable migration threshold on your resource so a low (or single) number of failures causes the resource to be migrated. Check out the configuration guide on the clusterlabs site, it has plenty of good examples. -- Regards, Oliver Hookins Anchor Systems _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
