On Mar 30, 2008, at 9:27 PM, Andreas Kurz wrote:
On Fri, Mar 28, 2008 at 6:02 PM, Andrew Beekhof <[EMAIL PROTECTED]>
wrote:
On Mar 28, 2008, at 3:14 PM, Dejan Muhamedagic wrote:
Hi,
On Fri, Mar 28, 2008 at 12:47:14PM +0000, David Thompson wrote:
Hello:
I have been asked to schedule a service managed by heartbeat to
restart every 24 hours (this is common practice in my company).
Interesting practice.
If I were to add a RESTART OPERATION to the CIB.XML for that
service with say an interval of 86400s, would that do the
trick? :-)
No. There's no restart operation. The only operation which can
recur is monitor anyway.
Actually thats wrong.
Recurring actions can have whatever name they like.
True the standard is "monitor", but there is nothing to stop David
from doing exactly what he's proposing - I believe it will work.
And because resource operations are serialized, there is no chance
that another recurring monitor will run at just the wrong time and
think the resource is failed.
How about simply stopping the resource "outside" of heartbeat e.g. via
cron and let the monitor action detect the "failure" and initiate a
restart of the resource? .... If resource_failure_stickiness is not
used to failover services....
sure, that would work too. but i think david's original idea is the
simplest
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
