copy pool to primary pool - reversing the flow

[foobar2devnull]

Hi guys,

We had to migrate our TSM servers to suit the companies new requirements but it seems we messed up somewhere...

When I ran a restore for node foobar, I got the following error:

ANR0565W Retrieve or restore failed for session 89598 for node foobar (WinNT). The storage volume AA0057L6 is inaccessible.
ANR1422W Read access denied for volume AA0057L6 - volume access mode="offsite".

This leads me to believe it was unable to find the data on the primary pool and had to look for it on the copy pool which is off-site. I called the tape back for the restore to complete but what about data redundancy?

Is there a way for me to:

(1) copy the data from the copy tape back to the primary file pool?
(2) Identify any other discrepancies between primary and copy pools?

Thanks!

 

[marclant]

When I ran a restore for node foobar, I got the following error:

ANR0565W Retrieve or restore failed for session 89598 for node foobar (WinNT). The storage volume AA0057L6 is inaccessible.
ANR1422W Read access denied for volume AA0057L6 - volume access mode="offsite".

This leads me to believe it was unable to find the data on the primary pool and had to look for it on the copy pool which is off-site. I called the tape back for the restore to complete but what about data redundancy?

I would check the activity log prior to those warnings to see what primary volume was not there. It is possible that the file in question was damaged on the primary volume, but it's also possible that the volume is unavailable. It always tries the primary first, so there should be more info in the activity log.

q ac sessnum=89598 begint=hh:mm endt=hh:mm begind=mm/dd/yyyy endd=mm/dd/yyyy

(1) copy the data from the copy tape back to the primary file pool?

Yes, it's called RESTORE STGPOOL.

(2) Identify any other discrepancies between primary and copy pools?

It's harder because it's not always a 1 to 1 relationship. You can backup multiple primary pools into a single copy pool. However, every time you do a backup stgpool, it copies all objects in the primary that do not exist in the copy pool. If the result is like below, then it's update to date:

tsm: SERVER1>backup stgpool vtl1-pool vtl1-copy
ANR2111W BACKUP STGPOOL: There is no data to process for VTL1-POOL.

 

[foobar2devnull]

Thank you for such a great answer!

I got nothing much from the activity log:

12/01/2015 14:25:14 ANR0406I Session 89598 started for node foobar
(WinNT) (Tcp/Ip foo (51858)). (SESSION:
89598)
12/01/2015 14:26:16 ANR8340I FILE volume /home/tsminst1/stg/file02/pf01_vol01F
mounted. (SESSION: 89598)
12/01/2015 14:26:16 ANR0510I Session 89598 opened input volume
/home/tsminst1/stg/file02/pf01_vol01F. (SESSION: 89598)
12/01/2015 14:26:16 ANR0514I Session 89598 closed volume
/home/tsminst1/stg/file02/pf01_vol01F. (SESSION: 89598)
12/01/2015 14:26:16 ANR1422W Read access denied for volume AA0057L6 - volume
access mode="offsite". (SESSION: 89598)
12/01/2015 14:26:16 ANR0565W Retrieve or restore failed for session 89598 for
node foobar (WinNT). The storage volume AA0057L6
is inaccessible. (SESSION: 89598)
[...]

I will however, make sure to run the restore stgpool command.

I run a backup stgpool on a daily basis so I am sure my active data is up to date. I also ran the following and found no problems with primary volumes:
> select volume_name, access from volumes where (access != 'READWRITE' and access != 'OFFSITE')

 

[marclant]

Check if there are damaged files:

query content /home/tsminst1/stg/file02/pf01_vol01F damaged=yes

Check for read or write errors:

query volume /home/tsminst1/stg/file02/pf01_vol01F f=d

 

[marclant]

I will however, make sure to run the restore stgpool command.

Only do that if you see you have damaged files.