TSM data restoration

[Hiten]

Hi,

I am having 1 node (OS: Windows 2008 R2 x64bit) in which daily selective backups happening for two folders. Backup is of small size of 2.5 GB & online. But when I try to restore the data its taking long time even to query the same data more than 4hrs & still not restoring the single file. I took another server in which I tried to restore one folder out of two (by using same client Node OPT file) but its not restoring.
schedule is set as below.
Policy Domain Name: WINDOWS_SERVER_POLICY
Schedule Name: SRVAPPS_BKP
Description: SRVAPPSBKP
Action: Selective
Subaction:
Options: -subdir=yes
Objects: C:\Folder1\FolderA\* C:\Folder2\FolderA\*
====================
Completion status details are as below
03/10/2015 09:28:27 --- SCHEDULEREC STATUS BEGIN
03/10/2015 09:28:27 Total number of objects inspected: 33,473
03/10/2015 09:28:27 Total number of objects backed up: 33,472
03/10/2015 09:28:27 Total number of objects updated: 0
03/10/2015 09:28:27 Total number of objects rebound: 0
03/10/2015 09:28:27 Total number of objects deleted: 0
03/10/2015 09:28:27 Total number of objects expired: 0
03/10/2015 09:28:27 Total number of objects failed: 1
03/10/2015 09:28:27 Total number of subfile objects: 0
03/10/2015 09:28:27 Total number of bytes inspected: 626.56 MB
03/10/2015 09:28:27 Total number of bytes transferred: 639.13 MB
03/10/2015 09:28:27 Data transfer time: 8.25 sec
03/10/2015 09:28:27 Network data transfer rate: 79,329.02 KB/sec
03/10/2015 09:28:27 Aggregate data transfer rate: 211.81 KB/sec
03/10/2015 09:28:27 Objects compressed by: 0%
03/10/2015 09:28:27 Total data reduction ratio: 0.00%
03/10/2015 09:28:27 Subfile objects reduced by: 0%
03/10/2015 09:28:27 Elapsed processing time: 00:51:29
03/10/2015 09:28:27 --- SCHEDULEREC STATUS END
03/10/2015 09:28:27 --- SCHEDULEREC OBJECT END QAMSAPPS_BKP 03/10/2015 08:20:00

Can you guide I select the PIT restore for latest date.

 

[Chimera]

What kind of media are you attempting to restore from e.g. Tape/VTL/Disk? Also whats happening in the server actlogs - it may be you have some form of contention delaying the restore process.

 

[marclant]

What command did you use to restore?
What's the state of the session during the restore (q session)?
Are there other processes running using tape (q proc & q mount)?

 

[Hiten]

Hi,
We have successfully restored the data but surprisingly it brought us full data till date form the beginning when the actual server comes for backup. We found these restored brought us all the Active /Inactive versions together which sum 18.08gb in place of 700mb.
We have taken step to create new DISK device class & started backup by creating new policy for storage pool.We found the backup time got reduced with same size.
Today when I started the restoration with Point In Time it started the restoration & brought data of 400mb from disk & suddenly started mounting the old medias where as actual backup still remains on disk. I could not understand the behavior.

 

[marclant]

Hi,
We have successfully restored the data but surprisingly it brought us full data till date form the beginning when the actual server comes for backup. We found these restored brought us all the Active /Inactive versions together which sum 18.08gb in place of 700mb.

What command did you use? It should not restore both active and inactive unless you specify that in your command, and it's not good practice because you have no control in the order in which they are restored and each version restored overwrites the previous. So at the end, you don't know if you have the right version. Using a point-in-time prevents that, looks like that's what you did after which is good.

We have taken step to create new DISK device class & started backup by creating new policy for storage pool.We found the backup time got reduced with same size.
Today when I started the restoration with Point In Time it started the restoration & brought data of 400mb from disk & suddenly started mounting the old medias where as actual backup still remains on disk. I could not understand the behavior.

TSM uses incremental forever, so even if all new backups go to disk, if you back FileA 3 years ago, if FileA still existed on the disk of the machine and is still 3 years old, this file is still considered active, therefore any incremental backup will skip it because TSM already has a copy. So if you have old active files, it's normal that it uses the original media it was backed up on.

 

[Hiten]

Thanks marclant. But my backup is not incremental its full. You can see the schedule details as stated above. So do you mean that it's running incremental daily & not full backup? I want to configure Active Data pool here so that only active data can remain in disk & inactive data move to media which can be made offsite. Can you guide if how can i make it?

 

[marclant]

Sorry, missed you did a full. Well, a few things could have happened. Your PITD is before the full or the disk pool got full and data migrated to disk.

For active data pool, check: http://www-01.ibm.com/support/knowl...m.ibm.itsm.srv.doc/c_stgpool_adp.html?lang=en Note that it will not help you with a PIT because some of the data would be inactive.