Le 22/01/2013 11:30, Josh Bowling a écrit :

> What if a new version of Pacemaker/Corosync is included in the updates? I'm
> sure updating it wouldn't cause any compatibility issues.

Nop it would not, the updated node should rejoin just fine, allowing you 
to roll-upgrade the other node so that in the end you have a perfectly 
valid and updated cluster.

Of course, I guess it depends what kind of gap in version we are talking 
about, anyway, read the changelogs maybe.

When updating pacemaker+corosync on a production cluster, I like to 
disable them at reboot (sudo chkconfig corosync off && sudo chkconfig 
pacemaker off) so that I can validate my corosync rings before starting 
pacemaker on the updated node (sudo corosync-objctl | fgrep member).

But that's just my way.

Cheers, and loose some stress with a lab environment.

-- 
Cheers,
Florian Crouzat
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to