On 10/13/11 09:42, Felix Frank wrote:
On 10/13/2011 09:38 AM, Bart Coninckx wrote:
OK, I see. Can my workaround be considered safe? Possible pitfalls I see
are upgrades that overwrite the DRBD resource agent script with the 1
second delay. Besides that I would expect this to be fine?
thx,
B.
Hi,
could it be possible to hack something up using a delay resource?
I'm thinking along the lines of "migrate" constraints, which work by
querying the uname of the respective peer.
If you could have pacemaker delay the promotion based on the peer's
uname, you could avoid the risk you mention (and personally, I feel this
is a mite less of a hack, although still bad enough).
Cheers,
Felix
It should produce a one second delay then, as tests showed that less
delay still produces a split brain.
Agreed that they both are quite ugly, but if it works with not further
consequences? ...
cheers,
B.
_______________________________________________
Pacemaker mailing list: Pacemaker@oss.clusterlabs.org
http://oss.clusterlabs.org/mailman/listinfo/pacemaker
Project Home: http://www.clusterlabs.org
Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
Bugs: http://developerbugs.linux-foundation.org/enter_bug.cgi?product=Pacemaker
