[ClusterLabs] DRBD split brain after Cluster node recovery

[Digimer]

On 2017-07-12 07:19 AM, ArekW wrote:
> Thanks for quick respond. I've found the fencing docs on Linbit and try
> to follow but I got an error regarding "disk" section:
> 
> drbd.d/global_common.conf:57: Parse error: 'resync-rate | c-plan-ahead |
> c-delay-target | c-fill-target | c-max-rate | c-min-rate | bitmap' expected,
>         but got 'fencing'
> 
> I set the config in global.cont but also in my storage.res file (not at
> the same time) all according to the manual like:
> disk {
>   fencing resource-only;
> }
> 
> I got the drbd running when I deleted the fencing phrase but then the
> fencing is not working.

First, make sure stonith is enabled (and tested!) in pacemaker. Then you
configure drbd to use 'fencing resource-and-stonith' (resource; alone
isn't enough) and set the 'fence-handler' to be the 'crm-fence-peer.sh'
script (and crm-unfence-peer.sh for after resync).

This is the only way to avoid split-brains.

-- 
Digimer
Papers and Projects: https://alteeve.com/w/
"I am, somehow, less interested in the weight and convolutions of
Einstein’s brain than in the near certainty that people of equal talent
have lived and died in cotton fields and sweatshops." - Stephen Jay Gould