Could anyone help with this error message??

  • MSG TEXT: EventID: 0x4000429C (17052) - Error: 14152, Severity: 10, State: 1

    Replication-Replication Distribution Subsystem: agent OURSERVER-

    Thanks,

     

    Jaybee.

  • It might be a deadlock issue. See the link below.

    http://forums.microsoft.com/MSDN/ShowPost.aspx?PostID=89701&SiteID=1

  • Hi,

    This means that your distribution agents were blocked somehow. Either by each other or - which is more frequent at us - by the Logreader Agent. Check the distribution.dbo.msrepl_transactions table if there's a gap at this time. You can make your agents log and/or set the ReadBatchThreshold/ReadBatchSize parameters of the Logreader Agent.

    If you vote for logging, be aware that it increases quite fast if you left it on the default level 2, and you'll get not too much if you use level 0.

    Erik

  • Thanks for the help, a few newbie-to-replication questions though:

    "Hi,

    This means that your distribution agents were blocked somehow. Either by each other"

    Do you mean the distribution agent at the other server? 

     

     or - which is more frequent at us - by the Logreader Agent."

    How would I tell which - and is this crucial?

     

    "Check the distribution.dbo.msrepl_transactions table if there's a gap at this time"

    At the time of the fault?  It was reported last week, and the table only contains entries for the last 48 hours.

    "You can make your agents log and/or set the ReadBatchThreshold/ReadBatchSize parameters of the Logreader Agent."

    How (and why) would I get the agents to log?  I found the ReadBatchSize, it's set at 500; what would be a better number?

    Sorry it's a lot of amateur questions!

     

    Jaybee.

     

  • Hi,

    First of all, if you're new to replication, I suggest you to read the first few chapters from SQL BOL (Books Online). It's really worth it.

    Do you mean the distribution agent at the other server? 

    The distribution agent can be either on the distribution or the subscriber server, depending on whether you use pull or push replication. Look for jobs in REPL-distribution category.

    How would I tell which - and is this crucial?

    (You need to know where was the error to resolve it.) This is a mysterious area for me as well, but if the distribution.dbo.msreplĂśtransactions table was written continuously then the distribution agent(s) was blocked. Be aware of a common issue: if you run tough queries against the subscription database you make the distribution agent unable to update the tables (this case use (nolock) clause if it fits your needs).

    At the time of the fault?

    Yes. You have a good backup from the DB, do you? Or you can wait for the next similar problem.

    How (and why) would I get the agents to log?

    SQL BOL - search - command prompt utilities - Replication Distribution/Log Reader Agent. See Output and OutputVerboseLevel. There are no better or worse values in general. If your replication keeps timing out with no good reason, you may try to lower the amount of data read at once.

    Hope it helped

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

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