[Linux-cluster] prevent ping-pong

[Pavlos Parissis]

2009/4/24 Lon Hohberger <lhh at redhat.com>

> On Fri, 2009-04-10 at 00:35 +0200, Pavlos Parissis wrote:
> > Hi,
> >
> > Is there a way to configure cluster to not ping-pong a service more
> > than N times?
> >
> > I have an application where sometimes cluster gets a no zero status on
> > the check and keeps failover back and forth the service.
> > I want the cluster to try only N times to fail over the service and
> > then disables it if it can't start it or gets a no zero status on
> > checks after N times.
>
> On most-current RHEL 4 + errata, you can do a 'max_restarts' combined
> with 'restart_expire_time' to achieve a maximum restart before failover,
> but there's no current limit to the # of restarts / failovers before a
> disable.


Thanks

Pavlos
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/linux-cluster/attachments/20090427/dec25623/attachment.htm>