[Linux-cluster] Fencing: Prevent rebooting halted node

[Nicolas Ecarnot]

Le 21/03/2012 12:23, emmanuel segura a écrit :
> Hello Nicolas

Hello Emmanuel,

> The first i can recommend it's use a quorum disk

I do use a quorum disk. I did not mention it as I thought this had 
nothing to do with my issue.

> and for your problem i
> know that it's called fencing-loop

Ok, I'll search on that subject.

> you got two choices
>
> 1:don't put the daemons cluster at boot time

In fact, I do want the cluster daemons to be ran at boot time. This is a 
good behaviour.
I just want dead nodes to remain dead, until manual intervention.

> 2:If you use a qdisk you can use clean_start="1" in the fence_daemon
> tag, this can be used if you are not using gfs or gfs2 or
> <cmantwo_node="1"expected_votes="1"/>

I am using gfs2...


To cope with my problem, I was looking whether it was possible to call 
my own script in cluster.conf? I don't know if this is permitted and 
what return code is expected?

-- 
Nicolas Ecarnot