On Jan 21, 2008, at 11:01 AM, Sylvain MALAVIEILLE wrote:
Thanks for your answers.
My feeling is that heartbeat seams to be an insteresting cluster
software, but I am asking me some questions about resources and
failover capabilities.
My constatation is that start and stop orders are not respected in
the traces.
Is it normal ?
correct - ordering information is not listed in the log messages
...
Is start / stop orders are logged ?
yes
Where start / stop orders are logged ?
What lines bring the information ?
oh, sorry, i missed the "orders" part.
no, they're not logged. the information you're after is in the
transition graph - logging it is considered too verbose.
Is there any way to link 2 resources in order to initiate a
failover only if the two resources have failed.
not really
It could be done making an OCF script in order to manage the two
resources (instead of one).
But in my opinion this kind of actions should be done by heartbeat
itself.
I think that the real question is :
Is it possible to configure failover conditions for resources ?
If yes, what are the heartbeat capabilities about that ?
I think this is a very important part of a cluster software.
On a resource failure on all nodes (the resource try to start on
all the cluster nodes before) is it possible to :
start failures are fatal - they prevent the resource from being
tried there again until the admin fixes the underlying problem and
uses crm_resource -C
Same question :
"Is it possible to configure failover conditions for resources ?"
not for start or stop failures. for everything else, yes.
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
