[Linux-cluster] 4 node cluster even split

Lon Hohberger lhh at redhat.com
Thu Apr 12 18:31:29 UTC 2007


On Tue, Apr 10, 2007 at 03:20:01PM +0300, Janne Peltonen wrote:
> 
> Now I wonder. Would it be simpler to set up a qdisk? That'd require
> some consideration of the heuristics to use. And the documentation
> appears sparse. Is this correct:
> 
>  -whichever partition of a partitioned cluster contains the qdisk
>  master, has all the qdisk votes,
> 
>  AND
> 
>  -qdisk master will always be in the partition that contains the
>  lowest-numbered cluster node that considers itself feasible to be in
>  the cluster (regardless of the opinion of the other nodes in that
>  partition)?

Actually, no... if that's in the FAQ, it's wrong.  Qdisk will elect a
master, which is generally the lowest node - however, that doesn't mean
that the master status is removed simply because of an even split.

If you have an even split with qdisk with incorrect heuristics (or no
heuristics at all with the updated test packages), the partition which
currently has the master will remain quorate.

> 
> In my scenario (failing rack), the qdisk master and the qdisk votes
> would find their way to alive partition and the partition would keep
> quorum. I think.

> 
> If I were to use a simple heuristic (such as pinging a router upstream),
> that wouldn't break anything in the failing-rack scenario. But what if
> there was a real split in the heartbeat network, with both halves of the
> cluster still seeing the upstream router (because there was no split in
> the 'outer' network)? If SAN access is kept, and both halves still see
> the quorum device, would the cluster halves be able to agree on the
> master?

Worst case is that they fence race.  With the current test packages,
the partition which contains the QDiskd master will (is supposed to?)
fence the other partition(s).

-- 
Lon Hohberger - Software Engineer - Red Hat, Inc.




More information about the Linux-cluster mailing list