Database Mirroring State

  • L' Eomot Inversé (9/21/2011)


    (which I believe is that replication continues as normal but with reduced replication latency - but I can't find a reference to back that up).

    I guess we need to test this scenario 🙂

    M&M

  • Thanks for the question. Replication is a weak area for me so I had to guess and guessed wrong.

    ---------------------------------------------------------------------
    Use Full Links:
    KB Article from Microsoft on how to ask a question on a Forum

  • Learned something new today. I missed the question because it was a little unclear I think.

    http://brittcluff.blogspot.com/

  • Thanks for the question. Replication and mirroring are topics I need to read up on more.

  • mohammed moinudheen (9/21/2011)


    L' Eomot Inversé (9/21/2011)


    (which I believe is that replication continues as normal but with reduced replication latency - but I can't find a reference to back that up).

    I guess we need to test this scenario 🙂

    Test away. I came accross this when I needed to take application consistant snapshots of my mirror server databases every 30 minutes for reporting purposes (Using NetApp SnapDrive article will be coming soon). What I found was that replication latency increased at during time the mirror was "suspended" or "paused". After a bit of research i found out that because the transactions where not "hardened" in the log that the replication log reader agent would not replicate the transactions. As this caused us quite a few issues as we have external facing customer websites accessing the subscriber databases a decision was made to enable trace flag 1448.

    http://support.microsoft.com/kb/937041

    For those who are still not sure on the details of the correct answer, see the cause section of the above link.

    Thanks

    Chris

    ------------------------
    I am the master of my fate:
    I am the captain of my soul.
    ------------------------
    Blog: http://sqlserver365.blogspot.co.uk/
    Twitter: @ckwmcgowan
    ------------------------
  • anthony.green (9/21/2011)


    Good question Chris

    Ant

    Thanks Ant, appreciate it.

    ------------------------
    I am the master of my fate:
    I am the captain of my soul.
    ------------------------
    Blog: http://sqlserver365.blogspot.co.uk/
    Twitter: @ckwmcgowan
    ------------------------
  • chris.mcgowan (9/22/2011)


    ... As this caused us quite a few issues as we have external facing customer websites accessing the subscriber databases a decision was made to enable trace flag 1448.

    http://support.microsoft.com/kb/937041

    For those who are still not sure on the details of the correct answer, see the cause section of the above link.

    Thanks

    Chris

    So the correct answer once more is: It depends.

    If you have trace flag 1448 enabled, transaction replication is not affected. If it's not, no transactions of the exposed publisher are replicated to the mirror. :Whistling:

    Regards,

    Michael

  • chris.mcgowan (9/22/2011)


    mohammed moinudheen (9/21/2011)


    L' Eomot Inversé (9/21/2011)


    (which I believe is that replication continues as normal but with reduced replication latency - but I can't find a reference to back that up).

    I guess we need to test this scenario 🙂

    Test away. I came accross this when I needed to take application consistant snapshots of my mirror server databases every 30 minutes for reporting purposes (Using NetApp SnapDrive article will be coming soon). What I found was that replication latency increased at during time the mirror was "suspended" or "paused". After a bit of research i found out that because the transactions where not "hardened" in the log that the replication log reader agent would not replicate the transactions. As this caused us quite a few issues as we have external facing customer websites accessing the subscriber databases a decision was made to enable trace flag 1448.

    http://support.microsoft.com/kb/937041

    For those who are still not sure on the details of the correct answer, see the cause section of the above link.

    Thanks

    Chris

    Are you sure that you were running mirroring in High-safety mode with automatic failover?

    According to the BoL page on Replication and Database Mirroring (and corresponding pages for earlier SQL Server versions) if the mirror is unavailable in that mode the log reader continues to send commands to the distributor, so it's clear that it doesn't wait for the log records to be hardened to the mirror. It's high performance where the behaviour you describe is encountered, according to the documentation (in high safety mode without automatic failure no activity is permitted in the master when the mirror is inaccessible, and I imagine that includes when mirroring is paused).

    Tom

  • Nice question, thanks.

    Need an answer? No, you need a question
    My blog at https://sqlkover.com.
    MCSE Business Intelligence - Microsoft Data Platform MVP

  • L' Eomot Inversé (9/22/2011)Are you sure that you were running mirroring in High-safety mode with automatic failover?

    According to the BoL page on Replication and Database Mirroring (and corresponding pages for earlier SQL Server versions) if the mirror is unavailable in that mode the log reader continues to send commands to the distributor, so it's clear that it doesn't wait for the log records to be hardened to the mirror. It's high performance where the behaviour you describe is encountered, according to the documentation (in high safety mode without automatic failure no activity is permitted in the master when the mirror is inaccessible, and I imagine that includes when mirroring is paused).

    That would have been my understanding as well, but haven't received any feedback to my post two days ago.

    Thanks,

    Michael

Viewing 10 posts - 16 through 24 (of 24 total)

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