Wednesday, March 16, 2016

WebSphere MQ Version 8, z/OS 8 byte log RBA - Converting your bootstraps.

In previous versions of WebSphere MQ, the LOG RBA was limited to 6 bytes, leaving enterprises to monitor the RBA so as not to exhaust the RBA and have the queue manager crash.  In order to reset the RBA, new bootstraps and logs needed to be allocated followed by running the CSQUTIL program using the RESETPAGE FORCE option.

Now, with version 8, the RBA has been extended to 8 bytes, virtually eliminating the possibility of running out of RBA log range.  Using a 6 byte RBA gave us a 256 terabyte log range. Now, with 8 bytes, we have 16 exabytes of log range, or 64K times more.  A queue manager running at 100MB per second, would take over 5000 years to run out of RBA log range.

So how do we convert to the 8 byte RBA.  If you are a standalone queue manager, you just need to be on MQ version 8 running in the NEWFUNC OPMODE.  If you are a queue manager running in a Queue Sharing Group (QSG), all of the members of the QSG need to be running version 8 using the NEWFUNC OPMODE.

Steps to convert your bootstrap to version 2 (8 byte):

1.) Stop queue manager cleanly.

2). Allocate a new bootstrap or set of bootstrap data sets using a different name than the currently used by the queue manager..

3). Run the CSQJUCNV conversion utility using the correct PARM.

      Standalone Queue Manager:
           PARM=('NOQSG')
      Queue Sharing Group Queue Manager:
           PARM=('INQSG,QSGroupName,DataSharingGroupName,DB2Member')

4). Rename the current bootstraps to another name...i.e. V1.

5). Rename the new V2 bootstraps to the names used by the queue manager.

6). Start the queue manager.

7). Verify that the new RBA range is shown in the JES.

    CSQJ034I <MQM1 CSQJW007 END OF LOG RBA RANGE IS FFFFFFFFFFFFFFFF

    The previous message showed the RBA as 0000FFFFFFFFFFFF

Below is some sample JCL using just a single bootstrap dataset.

Step 1: Allocate a new bootstrap with a different name:

//STEP1   EXEC PGM=IDCAMS
//SYSPRINT DD SYSOUT=*
//SYSIN    DD *
 DELETE (COM.MQM1.V8.BSDS01)   <=== NOTICE V8, New name
    DEFINE CLUSTER                          -
           (NAME(COM.MQM1.V8.BSDS01)        -
            VOLUMES(DCPSMS)                 -
            UNIQUE                          -
            SHAREOPTIONS(2 3) )             -
        DATA                                -
           (NAME(COM.MQM1.V8.BSDS01.DATA)   -
            RECORDS(850 60)                 -
            RECORDSIZE(4089 4089)           -
            CONTROLINTERVALSIZE(4096)       -
            FREESPACE(0 20)                 -
            KEYS(4 0) )                     -
       INDEX                                -
          (NAME(COM.MQM1.V8.BSDS01.INDEX)   -
           RECORDS(5 5)                     -
           CONTROLINTERVALSIZE(1024) )

            If you have 2 bootstraps....you would also allocate a V8.BSDS02

Step 2: Run the bootstrap conversion utility:

    //STEP3    EXEC PGM=CSQJUCNV,REGION=32M
  //*            PARM=('INQSG,++QSGNAME++,++DSGNAME++,++DB2SSID++') *QSG*
  //             PARM=('NOQSG')                                     *NOQSG*
  //SYSPRINT DD SYSOUT=*
  //SYSUT1   DD DSN=COM.MQM1.BSDS01,DISP=SHR        <== current BSDS01
  //*SYSUT2   DD DSN=COM.MQM1.BSDS02,DISP=SHR       <== current BSDS02 
  //SYSUT3   DD DSN=COM.MQM1.V8.BSDS01,DISP=OLD     <== new BSDS01
  //*SYSUT4   DD DSN=COM.MQM1.V8.BSDS02,DISP=OLD    <== new BSDS02


           In the above JCL, SYSUT2 and SYSUT 4 are commented out 
           since we are only doing a queue manager with a single BSDS. 
           If you are converting a queue manager running DUAL bootstraps, 
           then all 4 SYSUT# will be used.

Step 3: Rename the current bootstraps to a V71 name to move them out of the way:

    //STEP4     EXEC PGM=IDCAMS
  //SYSPRINT  DD   SYSOUT=*
  //SYSIN     DD   *
    ALTER 'COM.MQM1.BSDS01'       NEWNAME('COM.MQM1.V71.BSDS01')
    ALTER 'COM.MQM1.BSDS01.DATA'  NEWNAME('COM.MQM1.V71.BSDS01.DATA')
    ALTER 'COM.MQM1.BSDS01.INDEX' NEWNAME('COM.MQM1.V71.BSDS01.INDEX')
  //*

Step 4: Rename the new bootstraps (8 byte RBA) to the name the queue manager uses:

    //STEP4     EXEC PGM=IDCAMS
  //SYSPRINT  DD   SYSOUT=*
  //SYSIN     DD   *
    ALTER 'COM.MQM1.V8.BSDS01'       NEWNAME('COM.MQM1.BSDS01')
    ALTER 'COM.MQM1.V8.BSDS01.DATA'  NEWNAME('COM.MQM1.BSDS01.DATA')
    ALTER 'COM.MQM1.V8.BSDS01.INDEX' NEWNAME('COM.MQM1.BSDS01.INDEX')
  //*

Step 5: Start queue manager and check for new LOG RBA RANGE:

       CSQJ034I <MQM1 CSQJW007 END OF LOG RBA RANGE IS FFFFFFFFFFFFFFFF


That's it!!

Happy messaging.

7 comments:

  1. This comment has been removed by the author.

    ReplyDelete
    Replies
    1. Not sure what you mean "pole".

      Delete
    2. This comment has been removed by the author.

      Delete
  2. Sorry,
    Hello again,
    we want to open a new MQ Manager. Can we directly open 8 bytes?

    Thank you.

    ReplyDelete
  3. The new queue manager just has to be V8 running in NEWFUNC mode...and then you will need to allocate the new bootstraps...and run step 2 above to convert them to the 8 byte RBA.

    ReplyDelete
  4. Truely a very good article on how to handle the future technology. After reading your post,thanks for taking the time to discuss this, I feel happy about and I love learning more about this topic. keep sharing your information regularly for my future reference. This content creates a new hope and inspiration with in me. Thanks for sharing article like this. The way you have stated everything above is quite awesome. Keep blogging like this. Thanks.
    Hadoop Training in Chennai
    Hadoop Training
    Best Hadoop Training in Chennai
    Best Hadoop Training Institute in Chennai

    ReplyDelete