Hi.
I tested the control by fail-count.
I used two dummy resources.
If fail-count became 2, prmDummy1 set it to move to the standby node.
If fail-count became 1, prmDummy2 set it to move to the standby node.
1)I start two resources in an active node.
2)It caused a monitor error first.
Result, prmDummy2 moved.
3)A re-start error of prmDummy1 happened.
Result, prmDummy1 moved.
4)All resources moved to a standby node.
5)I cleared failcount to return a resource.
crm_failcount -D -r prmDummy1 -U nodeA
crm_failcount -D -r prmDummy2 -U nodeA
6)The resource that failed in a monitor returned to a nodeA.
But, the resource that failed in a restart did not return.
I think that I am caused by the fact that INFINITY is set by the value of
the resource that failed in a start.
----- lib/crm/pengine/unpack.c unpack_rsc_op ---
if(task_status_i == LRM_OP_NOTSUPPORTED
|| is_stop_action
|| safe_str_eq(task, CRMD_ACTION_START) ) {
crm_warn("Handling failed %s for %s on %s",
task, rsc->id, node->details->uname);
resource_location(rsc, node, -INFINITY,
"__dont_run__failed_stopstart__",
data_set);
}
In this situation, is there the method how it is easy to return the resource
that failed in a start to an original node?
(configuration)
Heartbeat ver2.0.8
node 2node
Resource Dummy
crm yes
Regard.
Yamauchi.
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
