Reclamation Automation

[eFinity]

Hi,

I wrote a script to automate the reclamation process on TSM running on AIX. Script queries the STGs that have volumes with PCTUtil below 10%.


Here is the script:

#!/bin/ksh
dsmadmc -id=USERID -pa=PWD -dataonly=yes "select distinct e1.STGPOOL_NAME from volumes e1, libvolumes e2 where e1.volume_name=e2.volume_name and e1.status='FULL' and e1.PCT_UTILIZED < 10 order by e1.STGPOOL_NAME desc" > /home/SCRIPTS/rcl.txt

for i in $(cat /home/SCRIPTS/rcl.txt | awk '{ print $1}')

do

echo "THIS WILL BE THE COMMAND: reclaim stg $i th=90"

dsmadmc -id=USERID -pa=PWD "reclaim stg $i th=90"

done



I also defined a client schedule for tsm server node for the script to run every 8 hours.

The problem with the script is if there is no STG returned from the sql command or there is a job for the returned STG going on, script generates error.

I hope it also works for you and also hope someone can help me with the error messages.

[BBB]

Why are you only reclaiming full volumes? And you should use the %reclaim value rather than %used.

And why not just set the reclamation thresholds to what you want all the time if thats what you want. Otherwise have a schedule update the stgpool thresholds at the time you want to run reclamation.

So not really sure what this script does that couldn't be done in a much simpler way.

[eFinity]

It's my choise. I have copy pool process and not enough drives. And since I only look for full volumes, PCT_Util = 100 - PCT_Reclaim.

[Samuel]

Me, i have define a script :
PARRALLEL
reclaim stg stgpool_name thres=%
reclaim stg stgpool_name2 thres=%
...

and a script to stop reclaim
upd stg stgpool_name thres=100
upd stg stgpool_name2 thres=100
...