Results 1 to 7 of 7
  1. #1
    Join Date
    Mar 2002
    Posts
    4

    Unanswered: DB2 in Clustered/Failover Environment

    Hello,

    I am in the process of setting up a failover environment for DB2. We have two DB servers, and a disk array.

    The desired result is that DB2 is live and active on one of the DB servers - this server mounts the disk array where the database is located. (this much I have setup and working no worries) The second DB server sits in the background in the event of failure of the other DB server, in which case it will manually be mounted to the disk array and talk to the relevant database.

    This is where I am hitting an issue.

    So my question is - how am I able to make the second DB server aware of the database and point it to its relevant path.

    I am presuming there is some method of cataloging the already created database.

    Thanks,
    Sarah.

  2. #2
    Join Date
    Jun 2002
    Posts
    2
    Sarah,
    We have the MS Cluster Server setup on our production database close to what you have describe.

    These are the environment setup:
    1. Microsoft Cluster Services on two NT servers.
    2. Create group disks that can be made active on either at Node A or Node B server
    3. Create database on group disks
    4. Create virtual IP addresses that linked to physical IP addresses of Node A and Node B.
    5. Have application server's odbc connection point to this virtual IP address.
    6. Remove Instance out from Node B
    7. Follow UDB admin reference on running the udb cluster support utilities.

    Good Luck

    Randy

  3. #3
    Join Date
    Aug 2001
    Location
    UK
    Posts
    4,650

    Re: DB2 in Clustered/Failover Environment

    Sarah,

    What platform are you on ?

    Cheers

    Sathyarm


    Originally posted by SarahAko
    Hello,

    I am in the process of setting up a failover environment for DB2. We have two DB servers, and a disk array.

    The desired result is that DB2 is live and active on one of the DB servers - this server mounts the disk array where the database is located. (this much I have setup and working no worries) The second DB server sits in the background in the event of failure of the other DB server, in which case it will manually be mounted to the disk array and talk to the relevant database.

    This is where I am hitting an issue.

    So my question is - how am I able to make the second DB server aware of the database and point it to its relevant path.

    I am presuming there is some method of cataloging the already created database.

    Thanks,
    Sarah.

  4. #4
    Join Date
    Mar 2002
    Posts
    4

    Re: Re: DB2 in Clustered/Failover Environment

    Hi,

    Running on Redhat 7.2.

    It looks like the best way for me to achieve what we are after is to have two databases existant on the disk array - each DB server pointing to one database. Upon writing to the "live" database, I will need to replicate the data across to the "standby" database.

    Has anyone attempted such a scenerio - I haven't had the need to replicate data via DB2 before - time to start reading I guess :\

    Thanks,
    Sarah.


    Originally posted by sathyaram_s
    Sarah,

    What platform are you on ?

    Cheers

    Sathyarm



  5. #5
    Join Date
    Aug 2001
    Location
    UK
    Posts
    4,650

    Re: Re: Re: DB2 in Clustered/Failover Environment

    Data Replication , if you are talking about the DB2 Replication feature, is NOT the right solution. DB2 Replication is meant to be Asynchronous .


    I have heard of a couple of Linux HA Solutions which support DB2 . I think you need to investigate about them.


    My understanding of Failover configuration with AIX DB2 is as follows :

    Have the instance owner home directory and all the other filesystems containing the data be accessible from both the live and failover systems. Create the instance owner ids with the same uid on both the systems(this is part of the basic failover setup). When the the failover system is to be used, stop the instance on the live system , mount the requisite file systems on the failover system and issue db2start .


    Should be something quiet similar on Linux .....

    I'd appreicate if you keep your progress posted to the group.

    Cheers


    Sathyaram




    Originally posted by SarahAko
    Hi,

    Running on Redhat 7.2.

    It looks like the best way for me to achieve what we are after is to have two databases existant on the disk array - each DB server pointing to one database. Upon writing to the "live" database, I will need to replicate the data across to the "standby" database.

    Has anyone attempted such a scenerio - I haven't had the need to replicate data via DB2 before - time to start reading I guess :\

    Thanks,
    Sarah.



  6. #6
    Join Date
    Mar 2002
    Posts
    4

    Re: Re: Re: Re: DB2 in Clustered/Failover Environment

    Thanks for your reply.

    The process you have described is how I envisaged doing it. However, when I spoke to IBM support they advised against it and suggested replication was required.

    With the process you have detailed - my concern still exists re: the db alias to be used. The live server references the db by an alias - the failover server is unable to reference the db by the same alias. Therefore unable to reference the db.

    Am I looking at this from an incorrect angle and getting confused?


    Originally posted by sathyaram_s
    Data Replication , if you are talking about the DB2 Replication feature, is NOT the right solution. DB2 Replication is meant to be Asynchronous .


    I have heard of a couple of Linux HA Solutions which support DB2 . I think you need to investigate about them.


    My understanding of Failover configuration with AIX DB2 is as follows :

    Have the instance owner home directory and all the other filesystems containing the data be accessible from both the live and failover systems. Create the instance owner ids with the same uid on both the systems(this is part of the basic failover setup). When the the failover system is to be used, stop the instance on the live system , mount the requisite file systems on the failover system and issue db2start .


    Should be something quiet similar on Linux .....

    I'd appreicate if you keep your progress posted to the group.

    Cheers


    Sathyaram





  7. #7
    Join Date
    Aug 2001
    Location
    UK
    Posts
    4,650

    Re: Re: Re: Re: Re: DB2 in Clustered/Failover Environment

    The following text in the Replication Guide is in favour of my comment on Replication.


    Maintaining a backup server (hot-site backup): Do not use asynchronous
    replication to maintain a backup server that can be accessed when your
    primary server is down. If the source (primary) server becomes unavailable,
    there is no way to guarantee that all updates were made to the target
    (backup) server. If you must maintain a backup server, consider using other
    tools or features. For example, for System/390 consider Peer-to-peer
    remote copy (PPRC) hardware feature or the extended recovery component
    (XRC).



    Originally posted by SarahAko
    Thanks for your reply.

    The process you have described is how I envisaged doing it. However, when I spoke to IBM support they advised against it and suggested replication was required.

    With the process you have detailed - my concern still exists re: the db alias to be used. The live server references the db by an alias - the failover server is unable to reference the db by the same alias. Therefore unable to reference the db.

    Am I looking at this from an incorrect angle and getting confused?



Posting Permissions

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