ADSM-L

Re: Autodelete of SQL backups

2000-03-21 13:03:01
Subject: Re: Autodelete of SQL backups
From: Thomas Denier <Thomas.Denier AT MAIL.TJU DOT EDU>
Date: Tue, 21 Mar 2000 13:03:01 -0500
> I've got TDP (version 1.1.2) installed on an NT (4.0 sp5) client
> (version 3.7.1.03) with 50 SQL databases.  Its been backing up for
> several months and last week I noticed it was taking up nearly half a tb
> of tape space - actually closer to 400 gb's.  So I ran the autodelete
> command, successfully, and 2072 of 2373 objects were deleted.  When I
> check the node with the query occupancy command, it still shows all the
> versions exist.  Will I need to wait for the tapes that the deleted
> backups reside on get reclaimed, or is something else going on?  I ran
> the autodelete last Thursday and expiration and reclamation have run
> several times since then.

I have worked with the ADSM Connect Agent for Microsoft SQL, which is
essentially an earlier release of TDP for that database software, and almost
certainly handles backup retention the same way. When the Connect Agent
decides that a backup has reached the end of its retention period, it tells
the ADSM server that the client file corresponding to the backup has been
deleted. It expects the ADSM server to delete the backup immediately. This is
a different behavior than one wants for file-oriented backups, which should
keep backups around for a while after the corresponding client files have been
deleted. In order to achieve this difference in behavior, Connect Agent backup
files should be directed to a management class with different retention
parameters than management classes used for file-oriented backups. I suspect
that your 'deleted' TDP backup files are in a management class set up for
file-oriented backups, and will be kept around for the length of time
specified for 'Retain Only Version' in the backup copy group for the
management class.
<Prev in Thread] Current Thread [Next in Thread>