I went ahead and did the work today, everything went OK. Here's the
process I followed:
Confirmed that our Operations group had a clean backup (just in case!)
Exported out the cluster resource settings, built some batch files
to re-create the disk resources and shares later.
I shut down the passive node of each cluster (to keep multiple
systems from accessing the same "drive")
Then I went down the line, took the file share resources off line
Took the physical disk resources off line
Took the cluster name and IP offline (we're keeping the cluster name
on the new server)
Deleted the physical disk resources from the old cluster.
On the CX700 I removed the LUN's from the storage group with the 32
bit servers, and added it to the storage group with the 64 bit
Set up the cluster name and IP on the new cluster
For each disk, I used disk manager to assign the appropriate
letter.
Ran the batch file to create each disk resource and bring it on
line
Ran the batch file to create each share
Brought the passive node up
Failed over and then back.
Everything is working good. The cluster resources take about 30 seconds to
fail over (210 shares, 8 "drives", 4+TB, something like 12 million files).
I actually did the move on the first couple of drives one at a time,
without using the batch files. Once I felt it was working OK, I did the
rest of the drives together.
"John Toner [MVP]" <> wrote in message
news:...
> Looks like you've pretty much got it covered. Only thing I'd add would be
> to make sure that the devices are removed from the storage group on the
> CX array before adding the LUNs to the storage group for the 64 bit
> hosts.
>
> Regards,
> John
>
> Visit my blog: http://msmvps.com/blogs/jtoner
>
> "Mike O." <> wrote in message
> news:%...
>> I'm not "copying" anything, we're not switching SANs, we're switching
>> the cluster servers. Our existing cluster is a two node cluster running
>> on 32 bit Windows 2003 R2 enterprise. The "drives" are storage LUN's on
>> an EMC CX700 array. What I've done is create a two node x64 Windows
>> 2003 R2 enterprise cluster (with a new quorum disk). The new cluster is
>> on the same SAN, so I want to move the LUN's with the user data from the
>> existing cluster to the new one, without having to do a backup/restore
>> of 4TB+ of data. Since I'm going from 32bit to 64bit, I couldn't just
>> add the new nodes and retire the old ones.
>>
>> I have a couple of test "drives (LUNS)" with some junk data on the
>> existing cluster. I took one drive resource offline, deleted the
>> cluster resource, then moved it from from the old storage group to the
>> new one on the CX700 array. I assigned the drive letter on the new
>> server, then manually created the cluster resource following MS
>> KB555312. The new drive shows up OK, and it appears all the data is OK,
>> but I just wanted to make sure I wasn't missing anything.
>>
>> I did something like this several years ago in a test environment, but
>> again, I just wanted to try to avoid any mishaps.
>>
>> Regarding the link to the blog, we're basically doing the opposite of
>> what he's talking about. We're keeping the same SAN, but moving to a
>> new cluster.
>
>