You are correct. If you want the ability to lose a majority of the nodes and
still maintain the cluster, you must use the "No Majority" quorum model.
Microsoft states that this isn't the recommended quorum model...but it's
seriously been working fine for many moons in NT/2000/2003 clusters, so it's
not as big a deal as they make it out to be. The reason why this is no
longer recommended is because the disk designated as the quorum is
considered a single point of failure. If this disk dies, you cluster also
dies.
The recommended quorum model for a 3-node cluster is Node Majority, but this
only allows you to lose a single node. If two of the three nodes fail, the
cluster will lose its quorum and shut down the cluster completely. You can
always "Force" the quorum at this point to get it back online on the single
node, but this will need to be done manually.
I would not recommend MNS+FSW or Node+Disk quorum models as this does not
buy you anything. You still can only lose one of these votes before losing
your quorum, so there's no real benefit for these quorum models.
Hope this helps.
Regards,
John
Visit my blog:
http://msmvps.com/blogs/jtoner
"Peter Brown" <> wrote in message
news:9421A220-72D0-494A-B04D-...
> We're setting up a 3 node 2008 file server cluster and I was wondering
what's
> the recommended Quorum type for 3 physical nodes?
>
> I've read about the MNS File share quorum, but if we use that, we've got 4
> electors and it needs to be odd numbered, to it'll recommend Node
Majority.
>
> So if I want to run it all on 1 node, my only option is to configure it
with
> No Majority; Disk Only, correct, or am I missing something?
>
>
> Thanks,
>
> Peter
>