[corosync] Auto ring recovery issue

Dan Frincu df.cluster at gmail.com
Wed Nov 16 07:25:13 GMT 2011


On Fri, Nov 11, 2011 at 6:15 AM,  <manish.gupta at ionidea.com> wrote:
> Hi,
>  I am using corosync-1.4.2,
>  cluster-glue1.0.7,
>  pacemaker-1.0.11
>  Problem.
>  1.I have Configure only one ring, When ring goes down, split brain is
> observed and when ring comes up split brain is not resolved.Means after
> ring comes up it is not detected automatically.
>  2. Same is observed with 2 rings. When both ring goes down and again
> comes up , Split brain remains continue.
>  3. When I have configured 2 ring and if one ring goes down , and comes
> back up , It is automatically detected.
>  Please can you let me know this issue is fixed or not in corosync1.4.2
> when all configured ring goes down.

You haven't mentioned how many nodes you are using. Or if you have any
other resources running in the cluster, such as DRBD, and the split
brain is detected at the DRBD level or otherwise.

Automatic ring recovery works starting with corosync 1.4, that means
whenever a ring fails due to network connectivity interruption, on
restoration of network connectivity, corosync will re-enable the
failed ring automatically.

If that is not the case for you then test by running corosync-cfgtool
-s before and after the network issues.

One more thing, how are you making the ring go down? ifconfig ethx
down by any chance?

If so please read on http://www.corosync.org/doku.php?id=faq:ifdown


> Regards
> Manish
> _______________________________________________
> discuss mailing list
> discuss at corosync.org
> http://lists.corosync.org/mailman/listinfo/discuss

Dan Frincu

More information about the discuss mailing list