IBM Support

Installing a DB2 fix pack for the IBM PureData System for Operational Analytics

News


Abstract

To install a DB2 fix pack for the IBM PureData System for Operational Analytics, complete the following procedure. The DB2 fix pack can be installed on only the core warehouse hosts (administration, data, and standby data hosts).

Content

Important: The DB2 fix pack must not be installed on the management host nor on the standby management host.

Before you begin:

Determine and take note of the current version of the core warehouse DB2 software installed on your system.

Download the version of the Universal Fix Pack and NLPACK that matches the current version of the DB2 software on your system in case you need to revert to the old DB2 copy level. If your current version of DB2 software is a special build and unavailable through Fix Central, contact IBM Support for assistance in obtaining the DB2 special build software.

Download the version of the Universal Fix Pack and NLPACK to which you want to update your current version of the DB2 software.

Click the following link to access the Fix Central website to download the appropriate packages:


About this task:

You are going to install the DB2 fix pack into the <DB2DIR> directory of the current DB2 installation. The DB2 fix pack installation will overwrite the current contents of the <DB2DIR> directory.

Note: This procedure can also be used to install a DB2 special build.

Procedure:

1. To a directory accessible by all of the core warehouse hosts, extract the files, which you downloaded earlier, from the Universal Fix Pack and the NLPACK to which you want to update your current version of the DB2 software. The /stage file system is mounted on all of the hosts on the system so that extraction to this directory can be done on the management host. The following steps use the /stage/download directory on the management host as the source of the extracted DB2 fix pack installation software.

2. Stop the core warehouse DB2 instance.


    a. Log in to the management host as root.

    b. Determine the current location of partition 0 by running the following command:


      hals

    c. Stop the core warehouse DB2 instance by running the following command:

      hastopdb2

    d. The hastopdb2 command returns the status of the core warehouse DB2 instance. Verify that all of the database partitions are stopped.

      Note: In the IBM PureData System for Operational Analytics environment, the core warehouse DB2 instance will be stopped with the hastopdb2 command, but all of the GPFS file systems will remain mounted.
3. Install the fix pack on all of the core warehouse hosts by running the following command, which is a single-line command that might appear on multiple lines here due to page formatting reasons:

    dsh -n ${BCUDB2ALL} '/stage/download/universal/installFixPack -f NOTSAMP -f install -b <DB2DIR> -c /stage/download/nlpack > /stage/logfiles/FixPack_`hostname`_`date +%Y%m%d%H%M%S`.log'

    a. Before proceeding further, carefully check each of the log files for error messages. If error messages are found, contact IBM Support for assistance to resolve the fix pack installation issues.

    b. Verify that all of the core warehouse hosts contain the updated copy of the DB2 software by running the following command:


      dsh -n ${BCUDB2ALL} '/usr/local/bin/db2ls'

    This step completes the installation of the DB2 fix pack.

4. Upgrade the core warehouse DB2 instance.

    a. Run the following command as root on the host where you found partition 0 was located prior to stopping the core warehouse DB2 instance:

      <DB2DIR>/instance/db2iupdt bcuaix

    b. Run the following command on the management host as root:

      dsh -n ${BCUDB2ALL} "<DB2DIR>/instance/dasupdt"

    c. Prevent the DB2 Fault Monitor Controller Utility (FMCU) from being launched after a reboot on any core warehouse host.

      i. To show on which core warehouse hosts the db2fmcd controller is set in the inittab file to launch after a reboot, run the following command on the management host as root:

        dsh -n ${BCUDB2ALL} 'grep -i db2 /etc/inittab'

        The following example output is a result of running the previous command:

        host04: fmc:2:respawn:/usr/IBM/dwe/db2/V10.1.0.3..2/bin/db2fmcd #DB2 Fault Monitor Coordinator


        host02: fmc:2:respawn:/usr/IBM/dwe/db2/V10.1.0.3..2/bin/db2fmcd #DB2 Fault Monitor Coordinator
        host05: fmc:2:respawn:/usr/IBM/dwe/db2/V10.1.0.3..2/bin/db2fmcd #DB2 Fault Monitor Coordinator
        host06: fmc:2:respawn:/usr/IBM/dwe/db2/V10.1.0.3..2/bin/db2fmcd #DB2 Fault Monitor Coordinator
        host07: fmc:2:respawn:/usr/IBM/dwe/db2/V10.1.0.3..2/bin/db2fmcd #DB2 Fault Monitor Coordinator

      ii. To remove the db2fmcd controller startup entries from the inittab file, run the following command on the management host as root:

        dsh -n ${BCUDB2ALL} "/<DB2DIR>/bin/db2fmcu -d"

    d. Verify that the FMCU is not running on any of the core warehouse hosts by running the following command:

      dsh -n ${BCUDB2ALL} 'ps -ef | grep -i db2fm'
    • If the FMCU is running on a core warehouse host, run the following command on that host as root:
    • dsh -n ${BCUDB2ALL} "/<DB2DIR>/bin/db2fm -i <instance_name> -D"

5. Start the core warehouse DB2 instance by running the following command on the management host as root:

    hastartdb2

6. You might need to complete the following steps after starting the core warehouse DB2 instance:

[{"Product":{"code":"SSH2TE","label":"PureData System for Operational Analytics"},"Business Unit":{"code":"BU059","label":"IBM Software w\/o TPS"},"Component":["Not Applicable","Not Applicable"],"Platform":[{"code":"PF002","label":"AIX"}],"Version":"1.0","Edition":"","Line of Business":{"code":"LOB10","label":"Data and AI"}}]

Document Information

Modified date:
17 October 2019

UID

swg21695144