[DRBD-user] disaster management on primary-primary setup

Lars Ellenberg lars.ellenberg at linbit.com
Mon Apr 3 14:31:46 CEST 2017

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.


On Fri, Mar 31, 2017 at 02:37:30PM +0200, Gregor Burck wrote:
> Hi,
> 
> I'm testing drbd on a debian sytem. (drbd 8.9.2)
> My setup is two nodes with a primary-primary setup with gfs2

Debian is (was) missing the necessary fencing integration stuff
for integrating DLM with pacemaker fencing.

Node loss, but no fencing -> blocked GFS2.
Nothing to do with DRBD yet.

> I mount the cluster resource in the local filesystem. (/dev/drbd0 on
> /clusterdata type gfs2)
> 
> When I kill one node (take the electircal wire) the still existing node
> can't access the files:
> 
> ls -l /clusterdata hang up, I can't kill the command, even from a other root
> account with kill -9
> 
> Is this a problem with drbd or maybe with gfs2?
> 
> How to bring the situation under control?

My educated guess:
fix DLM -> Pacemaker fencing integration.

keywords, iirc: dlm.conf, dlm_stonith

Still, you also want to use "fencing resource-and-stonith;" in DRBD,
and appropriate fencing handlers for your overall setup.

-- 
: Lars Ellenberg
: LINBIT | Keeping the Digital World Running
: DRBD -- Heartbeat -- Corosync -- Pacemaker

DRBD® and LINBIT® are registered trademarks of LINBIT
__
please don't Cc me, but send to list -- I'm subscribed



More information about the drbd-user mailing list