Post reply

Transactional Replication Failure

  • April 1, 2008 at 11:43 am

    #185923

    there is a transactional replication setup and when we try to run the snapshot agent we get the following error:

    The replication agent has not logged a progress message in 10 minutes. Can anyone help us in understanding the issue and solving it.

    regards

    Preethi

  • April 2, 2008 at 1:14 am

    #797030

    Hi

    similar post have been there at this link:

    http://www.sqlservercentral.com/Forums/Topic458759-291-1.aspx#bm459301

    Please read the answer posted by dthomas.

    Hope this might help you.

    "More Green More Oxygen !! Plant a tree today"

  • April 2, 2008 at 5:06 pm

    #797568

    I have had this problem several times. From my experience it has been too many agents running and we needed to increase the desktop heap setting in the registry. Usually you can tell this is the problem if you stop a few agents and you are able to run it without problems.

  • April 2, 2008 at 6:12 pm

    #797576

    Thanks Mikelyn, Preeti would you please try that and post us the update.

    That will be of great help.

    "More Green More Oxygen !! Plant a tree today"

  • April 2, 2008 at 6:17 pm

    #797577

    If that is it I can send you our registry change we made. I could probably dig up the Microsoft article link to. This change was after we upgraded to 2005 and worked directly with our Microsoft people.

  • April 14, 2008 at 11:42 am

    #802206

    1. local disks are fast

    2. repl does row-at-once (lotsa round trips)

    3. networks are slow

    4. people send huge # of rows in one transactions

    5. some subs are way over-indexed (but reporting may need more than OLTP in PROD pub)

    so it can take a while for the Distribution Agent (DA) to complete the initial outload and/or subsequent changes. Hence the "no response from server for 10 minutes" or similar, but this is not a fatal and the DA continues pushing/pulling the data and will eventually complete.

    However, getting such alerts is ugly so you can tinker with settings [agent templates], encourage smaller transactions, check index costs/benefits.

    If you are repl several pubs (eg Pub.A -> Sub1.AA, Pub.B -> Sub1.BB), you should normally use shared DA's as the bottleneck is likely to be LAN and best not hog RAM at the Distributor.

    you should use SSMS to generate scripts and rework to your needs [see my post other threads for details]

    HTH

    Dick

Viewing 6 posts - 1 through 5 (of 5 total)

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