OK....agreed.

This is my dilemma:

I'm building a simple 2 node cluster, node1 and node2.
HA works fine so far.
If I issue "network stop" on node1, HA jumps in and transfers resources to
node2.
Good so far.
However, when I issue "network start" on node1 HA grabs all the resources
back and wants to be preferred node again.
I want node1 to NEVER return to the cluster without manual intervention.

Is this possible?

>From the name of the STONITH suicide device it sounded like it would allow
node1 to simply kill itself and avoid returning to the cluster.

I am using 2 bcast interfaces (serial is out of the question) and IPFAIL.

Any Help?

Thanks

On 8/8/07, Andrew Beekhof <[EMAIL PROTECTED]> wrote:
>
> On 8/7/07, Michael Dengler <[EMAIL PROTECTED]> wrote:
> > Hi,
> >
> > I am trying to figure out how to use the STONITH suicide device but
> can't
> > seem to find ANY info it.
> >
> > Any help is greatly appreciated.
>
> A key part of the way the stonith subsystem was designed is the O in
> STONITH.
> (hint: O := other)
>
> suicide plugins really have only limited usefulness and most of the
> time, if a node is healthy enough to receive and complete a stonith
> request... then it probably doesn't need to be shot in the first
> place.
> _______________________________________________
> 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

Reply via email to