Results 1 to 5 of 5

Thread: HADR & powerHA

  1. #1
    Join Date
    Aug 2008
    Posts
    26

    Unanswered: HADR & powerHA

    Hi Folks,

    We are having DB2 V10.1 Fixpack 1 and OS AIX 7.1

    Now our company wants a failover design which could support across locations. So DBA team wants HADR, but OS team wants HACMP. Now our company wants to know can we club these two functionalities, where HA clustering (OS level) shared disk @ primary site & HADR @ two different locations.

    Now my DBA team comes with one strong question, in case of service node failure on HA cluster, HA scripts will restart DB2 instance on the standby node and this will cause HADR connection to be terminated with secondary location which is at a fair longer distance.

    So will it ever be a good idea to merge these two functionalities. Also please be noted, there is no idea to setup another HADR on the primary site, like cluster node 1 acting as HADR primary & cluster node 2 acting as HADR standby.

    Valid information would be helpful guys. We need a simple design thats all.

  2. #2
    Join Date
    Jul 2013
    Location
    Moscow, Russia
    Posts
    666
    Provided Answers: 55
    Hi,
    Quote Originally Posted by dharmaraj_ganesan View Post
    So DBA team wants HADR, but OS team wants HACMP. Now our company wants to know can we club these two functionalities, where HA clustering (OS level) shared disk @ primary site & HADR @ two different locations.
    I don't see any conflict here.
    If I got you correctly, you want to have shared disk cluster at Primary and HADR with remote Standby. Do you?
    If so, you can use HACMP or TSA for Primary. But you will not able to automate HADR roles switch. You will have to do this roles switch manually if you need it.
    Just set up a HADR standby like in case of simple Primary using VIP of Primary for hadr_remote_host on Standby.

    Quote Originally Posted by dharmaraj_ganesan View Post
    Now my DBA team comes with one strong question, in case of service node failure on HA cluster, HA scripts will restart DB2 instance on the standby node and this will cause HADR connection to be terminated with secondary location which is at a fair longer distance.
    Since you use VIP on Primary, new Primary and Standby should reintegrate without any problems.
    Regards,
    Mark.

  3. #3
    Join Date
    Aug 2008
    Posts
    26
    Thanks Mark for your response.

    Your below question is correct.

    "If I got you correctly, you want to have shared disk cluster at Primary and HADR with remote Standby. Do you?"

    Now i have another scenario for you in this case, in case of service node goes down & if new primary takes more time to get online will that cause potential timeout for HADR connection to break out? Or will standby be unaware of all these switch over and it will keep pinging on primary till any one node comes alive???

    My only aspect is to keep HADR link open all the time. at no cause this link should go down.

    If you see any other potential failovers in HADR link with this design, can you please let me know, so that we will look into that area also. As we are very new to this deisgn. Also let me know if there any best practices/thumb rule/known issues with this design so that we can be well prepared for the same.

  4. #4
    Join Date
    Jul 2013
    Location
    Moscow, Russia
    Posts
    666
    Provided Answers: 55
    It doesn't matter for Standby what Primary really is - clustered or non-clustered.
    When Standby realizes that Primary is down, Standby will keep trying to connect to Primary until it comes alive.
    I would suggest you to read this best practices book as well: DB2 High Availability Disaster Recovery
    Regards,
    Mark.

  5. #5
    Join Date
    Aug 2008
    Posts
    26
    Thanks a lot Mark for understanding my requirement and giving positive comments for my questionnaire. Sure will go through the best practices book as well.

Posting Permissions

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