Backup Performance Investigation

You may also like...

3 Responses

  1. David Mooney says:

    So what was the final verdict? Was there another process hitting the backup target disk on Sundays? I’ve had similar customer backup performance issues that I tried to blame on concurrent I/O but was never able to close the loop entirely.

    There’s also a weird situation with the stats from your “good” backup: 24 buffers going to MC#0 and 172 buffers going to MC#1. Those numbers should be roughly similar if the targets are similar. The stats for the “bad” backup are also slightly skewed but not nearly to the same extent.

    • Ember Crooks says:

      This is still an ongoing discussion with the client, so I don’t have a final resolution. Once a week having a long-running backup is not a severe enough problem to get a lot of attention. This client is also on the smaller side and doesn’t always understand that their DBAs are not also experts in the backup tool, so I suspect it may be a while until final resolution. I just hope we never have to do a restore at that time.

  2. Richard Middleton says:

    Thank you Ember!
    This article really helped me out 🙂

Leave a Reply

Your email address will not be published. Required fields are marked *