Results 1 to 8 of 8
  1. #1
    Join Date
    Oct 2013
    Posts
    3

    Unanswered: Commit Active for 15+ minutes?

    Hi there,

    DB 10.1 Fix Pack 2 running on RHEL x64 6.3. Running HADR to a standby node across a gigabit network.

    We've had two separate situations in which five transactions have been stuck in "Commit Active" for 15+ minutes.

    We ended up having to kill the application servers and force stop/start db2 to get service back up and running.

    It's difficult (impossible) to reproduce this in a controlled way - anyone have any immediate ideas as to why we're stuck in Commit Active this long?

    Thanks

  2. #2
    Join Date
    Jun 2003
    Location
    Toronto, Canada
    Posts
    5,516
    Provided Answers: 1
    What HADR mode are you using?
    ---
    "It does not work" is not a valid problem statement.

  3. #3
    Join Date
    Oct 2013
    Posts
    3
    We are using Near Synchronous mode...

    Are there any additional logs I should be checking for weirdness - or known issues you could point me in the direction of?

    Thank you!

  4. #4
    Join Date
    Jun 2003
    Location
    Toronto, Canada
    Posts
    5,516
    Provided Answers: 1
    Use db2pd -hadr to monitor HADR state. You may also want to show us all HADR-related database configuration parameters and registry variables.
    ---
    "It does not work" is not a valid problem statement.

  5. #5
    Join Date
    Dec 2007
    Location
    Richmond, VA
    Posts
    1,328
    Provided Answers: 5
    how much work had the 5 transactions done that they had to commit? were there any locks going on at the time? Is there anything in the diaglog on either primary or standby servers? was there a bottle neck in your network? You mention gigabit network, but is there other traffic there? lag?

  6. #6
    Join Date
    Oct 2013
    Posts
    3
    I will take a look at this tonight. Unfortunately, production system, cannot go poking around until a "quieter" time of the day. Reporting back soon.

  7. #7
    Join Date
    Aug 2008
    Location
    Toronto, Canada
    Posts
    2,367
    High availability disaster recovery (HADR) performance:
    IBM DB2 10.1 for Linux, UNIX, and Windows

    Monitor receive buffer on standby:

    "You can determine if the standbys log-receiving buffer is inadequate by using the db2pd command with the -hadr option or the MON_GET_HADR table function. If the value for the STANDBY_RECV_BUF_PERCENT field, which indicates the percentage of standby log receiving buffer that is being used, is close to 100, increase the DB2_HADR_BUF_SIZE setting"

  8. #8
    Join Date
    Dec 2007
    Location
    Richmond, VA
    Posts
    1,328
    Provided Answers: 5
    ???
    production system, cannot go poking around until a "quieter" time of the day
    ???
    If that's the case you have more problems than you know. I am on my production machine any time of day poking around ensuring no issues or investigating any that may crop up. Never have been able to understand those kinds of comments, at a former employer they had an oracle app that noone was allowed to touch during the day, else the users wouldn't be able to do their jobs. Someone should be on those(and apparently yours) machines though identifying the performance issue and taking action to get it resolved.

    Dave

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •