October 19, 2006 at 11:58 am
I have a DB that is 30GB. backing it up to a cluster resource of 100GB. The DB lives on the same drive as the backup. When the job completes, (it takes about 2 hours) several hours later, a backup server running Veritas with a large SAN, grabs the backup folder and backs it up to the SAN. The plan says to delete files over 1 day. I do not have enough room to keep but 1 backup.
The plan needs the available 30GB to create the next days backup. More than often, the plan does not delete the previous days backup and the job fails, and will continue to fail because their is not enough room to create another backup.
Is the "Delete Files over 1 day" a post part of the job and what can I do to guarantee there is enough room and only 1 file in the backup folder before the new job tries to run?
October 19, 2006 at 12:08 pm
I think you should look at using a backup compression program that will shorten the time and compress your backup file. Several vendors have them, including Quest, Idera and RedGate. In my experience, the resulting file will be about 25% of the original database size.
The Maintence plan deletes the backup after the new backup completes, so you are going to continue to have this problem unless you run a file system command to remove files before your maintenance plan runs.
Aunt Kathi Data Platform MVP
Author of Expert T-SQL Window Functions
Simple-Talk Editor
October 19, 2006 at 12:09 pm
It doesn't sound to me like the plan should be doing the deleting. I think I would have Veritas do it after it has grabbed the file. If the Veritas SW doesn't have the provision to do it, I'm sure that a .CMD file can be written to handle it. Obviously, some checking should be done to make sure the file actually WAS backup up to the SAN before it is deleted.
I had a problem recently at a new client's site where the maintenance plan was "failing", and there were no error messages as to why. They were doing an "all DBs" plan, including master, msdb and model... and only model was in full recovery mode... There was an informational message that master and msdb could not be backed up... "skipped"... and that was it. My transaction logs were not getting purged as a result of the failure... and there were over 30,000 of them on the disk.
The only change I made was to remove master and msdb from the maintenance plan, and put them in a separate maintenance plan that did not have transaction logs being backed up... eliminated the informational message, and the purging of log files started happening again.
Perhaps this is a solution for you; but, again, I think I would have Veritas do the delete.
David
Thank-you,
David Russell
Any Cloud, Any Database, Oracle since 1982
October 20, 2006 at 6:37 am
How about a separate, simple process (scheduled batch file, even) that moves (not copy) the file to the SAN?
October 20, 2006 at 2:30 pm
WHy not simplify the process even further. You have a SAN for the backups, why not backup directly to a SAN disk ?
RegardsRudy KomacsarSenior Database Administrator"Ave Caesar! - Morituri te salutamus."
Viewing 5 posts - 1 through 4 (of 4 total)
You must be logged in to reply to this topic. Login to reply