Re: Moving Cluster to new SAN
From: MC (MC_at_discussions.microsoft.com)
Date: 03/07/05
- Next message: Rodney R. Fournier [MVP]: "Re: Moving Cluster to new SAN"
- Previous message: Todd J Heron: "Re: Default gateway(s) on a multihomed NLB server in unicast mode"
- In reply to: John Toner [MVP]: "Re: Moving Cluster to new SAN"
- Next in thread: Rodney R. Fournier [MVP]: "Re: Moving Cluster to new SAN"
- Reply: Rodney R. Fournier [MVP]: "Re: Moving Cluster to new SAN"
- Reply: Mike Rosado [MSFT]: "Re: Moving Cluster to new SAN"
- Messages sorted by: [ date ] [ thread ]
Date: Mon, 7 Mar 2005 08:23:04 -0800
John:
Thanks for the response. I researched this after my post and I now see the
importance of the signature manipulation. Would you recommend doing this
via the ClusterRecovery or any other particular tool?
"John Toner [MVP]" wrote:
> Your plan (as written) will not work. MSCS is highly dependent upon the disk
> signatures and you will not get MSCS back up again without reinstalling from
> scratch if you cannot restore the signatures.
>
> If this were my cluster, I would:
>
> 1) Change the IP addresses of the cluster NICs/resources (if necessary).
> Follow directions in the following KB article for details:
> http://support.microsoft.com/kb/230356
> 2) Disable cluster disk driver on both nodes.
> 3) Move the servers to the new location...alter zoning, masking, etc.
> 4) Power up one node and restore the disk signatures to the new SAN volumes.
> I would reboot after making these changes.
> 5) Verify drive letter assignments are correct for the disks on BOTH nodes.
> (power down Node1 when verifying on Node2).
> 6) Restore data to the new cluster volumes
> 7) Re-enable cluster disk driver and reboot.
>
> No need to evict nodes IMO.
>
> Hope this helps.
>
> Regards,
> John
>
> "MC" <MC@discussions.microsoft.com> wrote in message
> news:0CD53D6C-21C5-42BC-B1A2-8B1D66D38D67@microsoft.com...
> > I need to move a Netbackup Master server on a WIN2K3 2-node cluster to a
> new
> > SAN in a physically different location. I will not have dual SAN/array
> > connectivity for the move.
> >
> > My plan is to:
> >
> > 1) Evict the idle node, and thus run the cluster as a single-node cluster
> > 2) Disable cluster services on each node
> > 3) Physically relocate the cluster servers, change IP addresses (hostnames
> > will remain the same) perform SAN zoning, lun-masking, etc., for the new
> > Quorum LUN and other shared disk resources
> > 4) Restart cluster services on single node, and bring the cluster up as a
> > single-node cluster
> > 5) Re-join the evicted node to the cluster
> >
> > Will I need to preserve the quorum db, and disk signatures under this
> > scenario? Or, should the cluster come back up with new logs, db, etc.
> > provided it sees the quorum drive as before ("Q: Drive)? Also, is there a
> > better way to do this?
> >
> > Thanks in advance for any assistance.
>
>
>
- Next message: Rodney R. Fournier [MVP]: "Re: Moving Cluster to new SAN"
- Previous message: Todd J Heron: "Re: Default gateway(s) on a multihomed NLB server in unicast mode"
- In reply to: John Toner [MVP]: "Re: Moving Cluster to new SAN"
- Next in thread: Rodney R. Fournier [MVP]: "Re: Moving Cluster to new SAN"
- Reply: Rodney R. Fournier [MVP]: "Re: Moving Cluster to new SAN"
- Reply: Mike Rosado [MSFT]: "Re: Moving Cluster to new SAN"
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|