[Linux-cluster] GFS on 3-node cluster corrupted after full network outage

[David Brieck Jr.]

On 12/7/06, Klaas <klaas at klaas.nl> wrote:
> Hi;
>
> We have Cluster software running because we need GFS.
>
> Lately, we had a 100% network outage and all three
> nodes kept on running (of course inquorate)
> and the GFS stayed available too.
>
> No fencing (I understand this is documented behaviour) but
> this way our GFS will get corrupted if the rest of the
> software on the nodes keeps accessing the GFS...
>
> Please advise,
> Klaas
>

http://sources.redhat.com/cluster/faq.html#fence_what

>From the FAQ:

#  Can't I just use my own watchdog or manual fencing?

No. Fencing is absolutely required in all production environments.
That's right. We do not support people using only watchdog timers
anymore.

Manual fencing is absolutely not supported in any production
environment, ever, under any circumstances.