auto_failback has no meaning in a crm cluster
have a look for documentation on default-resource-stickiness instead
On Jan 8, 2008, at 5:43 PM, Jason Price wrote:
I've been trying to set up a 2 node, heartbeat v2 cluster to present
Samba
shares.
For reasons I don't understand, the resource group I've created
REALLY wants
to run on my second node. If I start node1, without node2 running
at all,
the service will not start. If I start node2, the resource group
will start
on node2. If I stop heartbeat on node2, it will failover cleanly to
run the
process on node1, but as soon as node2 starts back up, it fails back
over to
node2.
/etc/ha.d/ha.cf looks like this:
-------------------------
#logfacility local0
use_logd yes
auto_failback off
bcast eth1
node node1.domain.com node2.domain.com
crm on
keepalive 1
initdead 33
-------------------------
and I've attached the two nodes cib.xml files. Both cib.xml files
were
generated from the hb_gui program
What I would like is for the resource group to be able to float
between the
nodes easily, and to not have a preference for either. I thought
setting
auto_failback to 'off' and fully restarting the cluster would do
this, but I
seem to be missing something.
Any insight would be appreciated.
--Jason
<cib1.xml><cib2.xml>_______________________________________________
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
