Reconcile space in container based pool

[Mita201]

I have two TSM servers, let's say server1 and server2
Both have container pools
I have set replication from server1 to server2, both with PROTECT and REPLICATE NODE.
Replication was successfull, but I have decided not to replicate it that way any more.
So I have decided to remove data from target server.
I have disabled node replication, and did protect stg container1 purgedata=all (took days to finish).
I can see my node not protected any more, my storage not replicated, but space consumed on disk on target server did not shrink at all, for filesystem where containers are.
How can I reclaim space on target server?

 

[marclant]

I'm not sure if you already know this or not, but it's important to understand the difference between protect stgpool and replicate node

PROTECT STGPOOL: As the name implies, it protects the storage pool, so it copies all the data in the storage pool to the target, regardless if node replication occurs. That permits you to recover the source storage pool should it get damaged.

REPLICATATE NODE: This is optionally done after a protect stgpool, this copies the metadata associated with every backup objects for that node so that the node can restore from the target in the event the source goes down. If a protect stgpool had not been done previously, it would also copy the data along with metadata.


If you are no longer protecting and replicating and you have broken the relationship between the two servers (both source and target are not standalones). This would mean that the data you have on the target (server2) is no longer needed. You could delete all the filespaces for the nodes and then delete the nodes that you don't need on the target.

 

[Mita201]

Hi,
Thanks for the answer.
Yes, I have deleted all the data both node-wise and storage-wise, but it didn't shrink containers on target pool on target server. I checked, reusedelay was one day.
Five days later, it suddenly deleted most of the data - shrunk containers, released space, with some unknown process, obviously very intensive, since it generated a lot of archive logs - it kicked automatic TSM DB backup.
I would like to have some control on it...

 

[marclant]

What's your reuse delay on the storage pool?

 

[Mita201]

I checked, reusedelay was one day.

 

[marclant]

I don't know why it would have waited 5 days then. Are there processes that ran that day that didn't run the previous 5 days?

 

[Mita201]

No, at least no process that I can find in actlog history.
As I have understood, there is constant monitoring process that takes care about containers, and you can see something about with q extentupdates, but it does not show much...