Page 1 of 2 12 LastLast
Results 1 to 15 of 24
  1. #1
    Join Date
    Jan 2013
    Posts
    11

    db2 9.7 active-active cluster

    Hi, can anyone knows how to setup a db2 9.7 active-active cluster with automatic failover on linux or aix? it will be helpful.

  2. #2
    Join Date
    Jan 2003
    Posts
    4,147
    I think they call it PureScale. There is plenty of documentation on it.

    Andy

  3. #3
    Join Date
    Jan 2013
    Posts
    11
    Hi Andy, thanks for your answer, but i think PureScale is available for db2 9.8 onwards.

    P.D. sorry for the bad english.

  4. #4
    Join Date
    Jan 2003
    Posts
    4,147
    DB2 V9.8 is just V9.7 with PureScale. They are basically the same thing.

    Andy

  5. #5
    Join Date
    Jan 2013
    Posts
    11
    well yeah, but i need to do it with V9.7 that is the problem

  6. #6
    Join Date
    Jan 2003
    Posts
    4,147
    PureScale is the only way I know of to do active-active clusters. If you have V9.7 and add PureScale you just move to V9.8. All you did was add PureScale. Why are you fixated on V9.7?

    Andy

  7. #7
    Join Date
    Jan 2013
    Posts
    11
    im not shure but i think is because of licensing problems and you mean that without the version 9.8 is impossible to configure an active-active cluster?

    I appreciate your answers.

  8. #8
    Join Date
    Jan 2003
    Posts
    4,147
    Define what YOU mean by active-active cluster.

    Andy

  9. #9
    Join Date
    Jan 2013
    Posts
    11
    I Mean both nodes needs to be operational, ie the load is distributed between the two nodes via a balancer and in the event that one node fails the entire load is directed towards the node that is still active.

    I hope I expressed myself well.

  10. #10
    Join Date
    Jan 2003
    Posts
    4,147
    That is PureScale. You may find a third party solution, but it would mean adapting the application to it. Without PureScale, IBM uses a shared nothing architecture. This means that each node in a cluster has only its own data, it is not shared between them (DPF). There is no way to do it that I know of without PureScale.

    Andy

  11. #11
    Join Date
    May 2003
    Location
    USA
    Posts
    5,737
    Quote Originally Posted by pgarciab View Post
    I Mean both nodes needs to be operational, ie the load is distributed between the two nodes via a balancer and in the event that one node fails the entire load is directed towards the node that is still active.

    I hope I expressed myself well.
    You can not have two servers (and two databases) accessing the same data on disk at the same time unless there is specialized software like PureScale (or Oracle RAC) to manage it. Otherwise there would a "split-brain" and no way to enforce lock management to ensure data integrity.

    DB2 ESE version has HADR, which has active-passive, and each server having its own copy of the data (not shared disk), but only one server is active at a time. One can also do active-passive without HADR using shared disk between the two servers, but that needs a cluster manager to move a virtual IP address to the active server (and do some other stuff with disk mounts).
    M. A. Feldman
    IBM Certified DBA on DB2 for Linux, UNIX, and Windows
    IBM Certified DBA on DB2 for z/OS and OS/390

  12. #12
    Join Date
    Jan 2009
    Location
    Zoetermeer, Holland
    Posts
    746
    Quote Originally Posted by Marcus_A View Post
    DB2 ESE version has HADR, which has active-passive, and each server having its own copy of the data (not shared disk), but only one server is active at a time.
    The 'passive' side of a HADR config can be active as read-only copy. When you activate that then you must pay for the passive licence as well.
    Somewhere between " too small" and " too large" lies the size that is just right.
    - Scott Hayes

  13. #13
    Join Date
    Jan 2013
    Posts
    5
    You may also consider checking DB2 Database Partitioning Feature. Normally DPF fits in OLAP systems -where data is partitioned across nodes- and pureScale fits in OLTP systems. DPF node can be configured for Failover using TSA, powerHA or MS Failover Clusters. Please check DB2 High Availability Options redbook for more details.

  14. #14
    Join Date
    May 2003
    Location
    USA
    Posts
    5,737
    Quote Originally Posted by amosaad View Post
    You may also consider checking DB2 Database Partitioning Feature. Normally DPF fits in OLAP systems -where data is partitioned across nodes- and pureScale fits in OLTP systems. DPF node can be configured for Failover using TSA, powerHA or MS Failover Clusters. Please check DB2 High Availability Options redbook for more details.
    DPF provides scalability, but not high availability.
    M. A. Feldman
    IBM Certified DBA on DB2 for Linux, UNIX, and Windows
    IBM Certified DBA on DB2 for z/OS and OS/390

  15. #15
    Join Date
    Jan 2013
    Posts
    11
    First of all let me thank you all for your answers, now I have more clear the issue of high availability of DB2. Now I wonder if anyone knows the steps to configure an active-passive cluster with automatic failover. I am new to DB2 so any response is helpful.

    Regards.

Tags for this Thread

Posting Permissions

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