[DRBD-user] [PATCH] dopd should notify when peer is dead (was "Refusing to be Primary while peer is not outdated" when peer is dead?)

Wolfram Schlich lists at wolfram.schlich.org
Tue Mar 4 09:37:55 CET 2008

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

* Lars Ellenberg <lars.ellenberg at linbit.com> [2008-02-28 15:56]:
> On Wed, Feb 27, 2008 at 10:22:01PM +0100, Wolfram Schlich wrote:
> > This is a really nasty bug and renders the cluster kinda useless,
> > because the failover node won't run the resources (that depend on
> > drbd) when the active node dies. Is this assumption correct?
> > 
> > If yes, I should stop using that broken dopd ASAP... :>
> if you use heartbeat 2.1.3, and it dopd does causes these "handler
> broken, returned 20" kernel messages, then yes, you happen to use the
> broken dopd, and should stop doing so until we release a (hot?)fix.

Ok. So, is it still hot? :) What's the plan?
I'd like to include the patch in my Gentoo Heartbeat packages.

> heartbeat 2.1.2 dopd is ok in this regard (but has other shortcomings).

Well, I guess one doesn't really want to go back to 2.1.2.

> one possible hotfix would be:
> perl -i -pe 's{drbd-peer-outdater}{$&.sh}' drbd.conf
> cat > /usr/lib/heartbeat/drbd-peer-outdater.sh <<___
> #!/bin/sh
> /usr/lib/heartbeat/drbd-peer-outdater "$@"
> ex=$?
> if [ $ex = 20 ]; then
> 	ex=5
> fi
> exit $ex
> ___
> this is not strictly correct.
> but it makes it work as good as it gets right now.

Ok. I added that script to my Gentoo Heartbeat packages and added
a big fat warning for the users. Better than nothing.
Wolfram Schlich <wschlich at gentoo.org>
Gentoo Linux * http://dev.gentoo.org/~wschlich/

More information about the drbd-user mailing list