On 09/25/2012 10:32 AM, Zoltan Forray wrote:
> Thanks for the response.
>
> In reality, we are trying to move the existing backups to another TSM node
> that is going to take over the NFS mountpoint once this server
> is twilighted so we wont have to re-backup this mountpoint.
>

It's still not clear whether you're talking about backing up on
something that is _mounting_ the NFS directory, or something that is
_exporting_ the share.

If you've got LUNs exported to a Solaris server, on which filesystems
are written, I'm not sure how you'd get them to be read by (e.g.) a
Linux server.  Does ZFS Just Do This?  That'd be the obstacle that would
worry me.

Presuming you can get across the hurdle of mounting the raw disk with
exactly the same path, and presuming that Solaris and your linux of
choice treat any extended attributes sanely, you can bring it down to
the same problem:  all the filespaces associated with node X must move
with node X.  You can rename node SOLARIS_SERVER_1  to be
ABSTRACT_GROUP_OF_NFS_SHARES_1, deprecate /, /opt, user, etc.. from that
node, and soldier on.

But you won't be able to split off /exports/one_file_system from
/exports/the_other_file_system unless you do a filesystem export/import
to a different node name.



- Allen S. Rout