Try:
1)
adsm> select volume_name from volumeusage where node_name='MYNODE'
adsm> select volume_name from volumeusage where node_name='MYNODE'
output will be list of volumes
2) create a collocated temporary storage pool TEMP_COL
3) do a move data from each volume obtained in 1) to stg=TEMP_COL
4) for each volume created in TEMP_COL, do
adsm> q cont <vol> count=10
adsm> q cont <vol> count=-10
to confirm what node has data on each volume
5) do a move data for each volume of nodes that you don't need
collocated back to the original stgpool
6) at the end, you've left all of MYNODE's data in TEMP_COL

Note: if you've separated archive and backup data, restrict the volumes
that you operate on to those in the backup pool, so as not to waste time
on moving archive data.

Sound like a lot of work? Yes.  Wouldn't it be nice to have a command
like:
move data <vol> stg=<stgpool> node=MYNODE type=backup

I think so.  I've proposed it here before. :-)

Cheers..
Bruce
Brazner, Bob wrote:
>
> We use non-collocated storage pools.  If we know ahead of time that we need
> to perform a full restore of a node (be it a desktop PC or server), is
> there a way to co-locate just that one node prior to the restore?  The idea
> being this would minimize the number of tape mounts during the restore,
> thereby getting it done in the least amount of time.