[ClusterLabs] pacemaker self stonith

[Klaus Wenninger]

On 11/30/2017 11:41 AM, Hauke Homburg wrote:
> Hallo List,
>
> I am searching für a possibility to stonith a pacemaker node himself.
>
> The Reason is ich need to check of the pacemaker noch can reach the
> network outside the local network. Because of network outage.
>
> I can't connect to an ILO interface or so.

>From what I get a possibility could be using ping-resource.
Per default that would just signal the outcome of the ping-observation
via an attribute but you can configure it to actually fail the
resource if the ping-target isn't reachable.
If you then set on-fail=fence for operations start & monitor and
provided you have proper fencing in place that could already
be what you are searching for.
Wouldn't do that outside of pacemaker probably ...

Regards,
Klaus

>
> I consider a bash script:
>
> if [ ! print -c 1 8.8.8.8] then $Hardware_reset ; fi
>
> this in the crontab and run every minute.
>
> Thanks für Help
>
> Hauke
>
>
>