deteriorating backup times

  • Just recently my backup times have increased dramaticaly.

    I have a cluster which houses about 40 dbs. The usual full backup time is about 10 minutes. Yesterday that time increased to 120 minutes!

    Any ideas as to what might be happening?

    I will look at the job log to see what's up.

    thanks,

    warren

  • There are too many things that could affect this.  I would advise that you see if this is a trend before assuming anything dire.  If the backups are still sluggish today and tomorrow, you've got a problem.

    My first guess would be that something was competing with the backup job, either within the DBMS or on the server, or on the network if you're writing the backups to a share on a different server.  Antivirus programs running on the server can set the stage for a real clobbering time.

    If something was malfunctioning, again, network problems could be whacking you.

    I guess the most obvious question to have started with would have been, did your amount of data increase by a factor of twelve all of a sudden?  🙂  I'll assume it didn't.

    If your backups actually completed their tasks with a good return code, the SQL logs probably won't complain.

    I'm sure some of the others lurking in these here parts have better ideas.

  • there is nothing competing with the backups, and i'm not copying the backups to a network folder from the cluster.

    The Full backups have a good return code.

    Also, when the transaction logs overlap the full backups they are also taking a long time (from 5 minutes to 30 minutes), and increasing the CPU to 95-100%+ for 10+ minutes.

    When the transaction logs are run by themselves, time is about 5 mintues.

    Also i've just noticed that some backups take about 2-3 minutes for some dbs. when i look at the properties it says that they have 0.00mb Space available.

  • I would defintely make sure the transaction log dumps do NOT occur durring database backup times as this will cause problems.

    [when i look at the properties it says that they have 0.00mb Space available.]

     Is this the properties of the database itself? If so, that only means that there is no free space in the database and the database will expand the next time it needs space. If it is on your backup file, you are not getting good backups.

    I would also check to make sure the server itself is not getting backed up to tape at the time you are running the DB backups.

  • just one question regarding cpu usage.

    should the cpu be above 90% when the transaction logs are being backed up?

Viewing 5 posts - 1 through 4 (of 4 total)

You must be logged in to reply to this topic. Login to reply