TSM Management Class Retention

[tishimself]

Can someone tell me in layman's terms what is the difference between setting a management class to 0 or unlimted is?

Thanks

 

[LED888]

If the mgmtclass have a retention of 0 (zero) - No files are kept.

If mgmtclass have a retention of NoLimit - Keeping ALL versions. If a file changes daily ALL the versions will be kept. Database will grow, and buy a lot of scratch tapes.

Good Luck,
Sias

 

[tishimself]

Thank You LED! So to clarify... If I have an application that archives files to the TSM via an API which will only send te file one time and will also send force delete requests to the TSM then the Management class should be set for No limit. And if it is set to 0 it won't save any of the calls we push?

 

[LED888]

Archives are not the same as a backup.

An archive is a selective backup with a very long retention setting (default 365 days).

A selective backup is were the TSM Client backup the
file(s) regardless if the file have changed since the last backup.

Since there is a mention of the TSM API, are we using a TDP product to backup Oracle, SQL, or Domino?
TDP for Oracle, the amount of data that is being kept is controlled by RMan.

NoLimit means ALL versions will be kept nothing will be expire. There are some cavates.

Version Data Exist (how many) works with Retain Only version (how long).

If we set Version Data Exist to "NoLimit" and Retain Only also to "NoLimit." ALL versions of the file(s) will be kept forever.

If Version Data Exist is set to "NoLimit" and Retain Only is set to 30.
An active version is the most current backup of the file.
An inactive version is the previous version of the file.
An active file becomes an inactive version when the file on the work station have changed since the last backup.
Active version does not expire unless the following is set:
Version Data Exist: 0
Or
Retain Extra: 0

When an active version of the file becomes an inactive version, this file will be kept for 30 days.
If the file changes every day, by the end of the month would expect to see 30 versions of the file. Since Version Data Exist is set to "NoLimit" but we have Retain Extra Vesion set to 30. The inactive version will be kept for 30 days.

Correct, if the retention is set to zero nothing is being kept.

Good Luck,
Sias

 

[mbourne2006]

Hi Sias

Are you sure about the whole "archive is a special type of backup" statement? I'm not. A backup operation binds to the backup copygroup of the default management class, unless another class is explicitly stated at the client. The backup copygroup contains the retention parameters that you described. Strictly speaking, the "active" version of a file *will* expire (well, become inactive anyway) if the source file on the client has been deleted and a backup operation is performed. Otherwise, we'd be in the situation where the TSM server retains the most recent copy of all backups, whether files exist or not.

An archive, on the other hand, binds to the archive copygroup in the default management class, unless otherwise stated at the client. The only retention parameter here is the "RETVER" parameter that determines how many days an archive file should be kept. The other big difference is the assignment of a textual description to the archive package - which allows us to add files to an archive "package" during subsequent archive operations.

I guess one question that needs to be answered here is does the API client in question perform a backup or an archive? If archive, then no special changes need to be made to the node definition on the server to allow the API client to delete the existing archives. If the operation is a backup, however, then the node needs to be updated to have permission to perform deletions of its own backups. Something like "UPDATE NODE $NODENAME BACKDEL=YES" I think.

HTH



Kind Regards,
Matthew Bourne
"Find a job you love and never do a day's work in your life.

 

[LED888]

Matt, your correct that when the TSM Client perform a backup the data is bound to the backup copy group were as the archive data are bound to the archive copy group.

The reason why I mention that an archive is simply a selective backup but the data is kept for long term.
Just trying not to be technical.

With the backup function the TSM Client perform either an incremental backup or a selective backup.
Incremental - The file is backed up only if the file have changed since the last backup.
Selective - The file is backed up regardless if the file changed or not.

With the archive function it stores files unconditionally.
What I mean by unconditionally, is that there are no
version limit and the file(s) will be retain for a defined time period regardless of whether they changed since the last archive and/or are deleted from the workstation.

In terms of does the API bind the data to the backup copygroup or to the arhive copy group.
This depends on if we are calling a backup function call or an archive function call.
If we are using the API for the TDP to access the TSM Server. The data is sent as an archive object but it is bound to the backup copy group rather than the to archive copy group.

Good Luck
Sias