locked
Mirroring Connection Time Out RRS feed

  • Question

  •  

    Hi,

    I have Mirroring setup for 3 databases without a witness.

    The database mirroring monitor shows me that both Principal and Mirror are "Synchronized".

    SSMS also shows that the both instances are "Synchronized".

     

    However, the following error message appears couple of times in the event viewer:

    The mirroring connection to "TCP://abc.xyz.com:5022" has timed out for database "DB" after 10 seconds without a response.  Check the service and network connections.

     

    Also came across this article.

    http://support.microsoft.com/kb/947462

     

    Should i be worried?

     

     

    thanks 

     

     

    Wednesday, April 2, 2008 4:22 PM

Answers

  • Hi Rookie,

     

    You have set heartbeat time between principal and mirror database as 10 seconds and hence for every 10 seconds you principal will ping the mirror server and if its failed to connect you will receive the error mentioned above.

     

    Can you check data consistency between principal and mirror server by taking a snapshot on the mirrored db on mirror server and compare. If  everything is fine then you can change the heartbeat time to 60 sec using the below query in principal server on mirrored db's and check whether the error its repeating again.

     

    Code Snippet
    ALTER
    DATABASE dbname SET PARTNER TIMEOUT 60

     

     

     

    If the above fails then partner is not able to connect mirror server, check whether the ports used for mirroring is open in firewall or use telnet to ping it.

     

    Thursday, April 3, 2008 3:42 AM

All replies

  • Hi Rookie,

     

    You have set heartbeat time between principal and mirror database as 10 seconds and hence for every 10 seconds you principal will ping the mirror server and if its failed to connect you will receive the error mentioned above.

     

    Can you check data consistency between principal and mirror server by taking a snapshot on the mirrored db on mirror server and compare. If  everything is fine then you can change the heartbeat time to 60 sec using the below query in principal server on mirrored db's and check whether the error its repeating again.

     

    Code Snippet
    ALTER
    DATABASE dbname SET PARTNER TIMEOUT 60

     

     

     

    If the above fails then partner is not able to connect mirror server, check whether the ports used for mirroring is open in firewall or use telnet to ping it.

     

    Thursday, April 3, 2008 3:42 AM
  • Thanks Vidya.

    The snapshot did verify that the data is consistent.

    Changing the hearbeat to 60 sec fixed the problem.

     

    Thanks.

     

    Friday, April 4, 2008 7:14 PM
  • Hi there,


    I have the same problem with the Timeout (My Timeout is 60 seconds), i configured mirroring through a VPN to test the speed of the comunication because we are thinking to use it for DRP(Costa Rica to Panama). My big concern is the speed how the SQL sends the packages to the destination. For Example. My VPN can reach 9Megabits/sec of speed with a copy of files, but with the mirroring it can not take more than 1.5mbps . How can i improve this rate?.  also, we realized after we run the Index Maintenance  the Log grows at leas an extra 5GB and it hits the mirroring send rate over the VPN greatly (it adds hours)


    I appreciate any help or guide that you can give us.



    My Software: Windows Server 2003 R2, SP2,SQL 2005  SP3


    Wednesday, April 1, 2009 5:10 PM
  • HI Carlos,

    Have U resolved your problem.. If Yes, can U share the resolution, because I am facing similar problem.

    Wednesday, December 30, 2009 12:29 PM
  • I have also same problem..
    "SQLSERVER DBA" "INDIA"
    Wednesday, May 4, 2011 10:31 AM
  • Hi, Rookie, How did you take the snapshot ? I have a similar problem and i need to do likewise to see if it will solve the issue.

     

    Thanks.

    Friday, September 23, 2011 2:10 PM
  • Hi, How do you take the snapshot ?
    Friday, September 23, 2011 3:50 PM
  • when you implement mirroring you have to make sure the drive you place the transaction logs can support the additional workload. Make sure the disk waits for the transaction logs is under 20ms. If this is not the case you might want to setup 4 drives in a raid 10 to support the additional workload on the transaction log the mirroring causes. Also make sure you are not running a dbcc shrink file for the transaction log when you back it up. This cas also cause issues. I have seen the error before and fixed it by moving to faster disk and stopping shrink file.
    Thursday, September 13, 2012 7:41 PM