data size for each volume owned by a node

[sra006]

Hi,
I am trying to consolidate the data per node ,
initially we had a collocated pool( by node) , but lack of empty volumes negated this to the point that it is almost like a non-collocated pool. I can count the number of volumes per node in the pool(via sql), and use
q nodeddata "node" to see the amount of the data per volume, but the output is fixed and I need to have an
SQL to total the data amount per volumes.
With this I would be able to script to move data around ( consolidating smaller amounts, without moving all the volumes by move nodeddata) .
Does anyone came across this issue? Do you know which table(s) can have this info?
Thank you for checking,
Alex

 

[inthesun]

Hi Alex,

Here is an example of using the SQL sum function:

tsm: SERVER1> SELECT node_name, filespace_name, SUM(logical_mb) AS "Total MB" FROM occupancy WHERE node_name='NAME' GROUP BY node_name, filespace_name ORDER BY "Total MB" DESC

NODE_NAME FILESPACE_NAME Total MB
------------ ------------------ ----------------
NAME /db2archivelogs 219588.48
NAME /db2offlinebackup 76585.49
NAME /opt/sysadm 40167.95

If that is not helpful, have you attempted to do TSM collocation by Group? That way you can have high restore priority nodes mixed with nodes that are not needed to be restored as quickly (low restore priority). Schedule the backups of the high priority nodes first. Once done, then start the backup schedule for the lower priority nodes that will be able to use the same volume as the first scheduled nodes.
Here is a link to the define collocate group (it has links to the other needed commands)
www.ibm.com/support/knowledgecenter/SSGSG7_7.1.3/srv.reference/r_cmd_collocgroup_define.html