Upgrading IBM Storage Scale nodes that run on AIX

Use these steps to upgrade IBM Storage Scale nodes that run on AIX®.

1. On the node that you are upgrading, stop all user and applications activities in the IBM Storage Scale file systems.
2. Back up your file systems.
   Important: Creating a backup of your file systems is crucial to help ensure that your data is protected.
3. If file systems are mounted on the node, issue the mmumount command to unmount them.
   Warning: Do not use the -f option of the mmumount command.
4. Shut down IBM Storage Scale on the node.
   1. Issue the following command to shut down IBM Storage Scale.

      mmshutdown -a

      Note: Postinstallation scripts of an IBM Storage Scale upgrade handle mmccrmonitor and other mm processes, so you do not need to close them.
   2. Corroborate that IBM Storage Scale is turned off.

      mmgetstate -aL

      In the output, the GPFS status column must report down.

5. If you do not plan to restart the nodes after the upgrade, verify that the GPFS daemon is stopped and that the kernel extensions are unloaded.

   mmfsenv -u

   Note:

   • The mmfsenv -u command is normally used with assistance from support or for cases where its usage is documented such as in this scenario.
   • The mmfsenv -u command checks only the local node from which it is run, not all the nodes in the cluster.

   If the kernel extensions are unloaded, you get an output similar to the following example:

   mmfsenv -u
   GPFS: 6027-501 /usr/lpp/mmfs/bin/mmfskxunload: module /usr/lpp/mmfs/bin/mmfs unloaded

   If the kernel extensions are unloaded, you can continue with the upgrade and no restart of the node is necessary to be able to use GPFS.

   If the kernel extensions are still loaded, you get a Device busy error, similar to the following example.

   mmfsenv -u
   sysconfig(SYS_CFGKMOD-term): Device busy

   If the kernel extensions are still loaded, you can continue with the upgrade. However, to be able to use GPFS, you must wait until the node is restarted or the error is cleared.

   Troubleshooting: The Device busy error occurs when a process or an open file is keeping GPFS busy, thus preventing the kernel driver from unloading. The solution is to wait until after the installation is complete and then restart the node by using shutdown or reboot commands. For more information about these commands, see shutdown command and reboot or fastboot command in IBM AIX documentation. Alternatively, if you can identify the process that is keeping the kernel extensions busy, stop or end that process and then retry mmfsenv -u.
6. Download the installation package for the target version of IBM Storage Scale licensed programs package from IBM Fix Central.

   The package name is Storage_Scale_<edition>-<target_version>-<architecture>-AIX-update.tgz. For example: Storage_Storage_Scale_Advanced-5.2.3.7-ppc64-AIX-update.tgz.

7. Prepare the files for the upgrade.
   1. Copy the installation images to a temporary location.

      # cd /tmp/scale_<target_version>
      # ls -ltr
      total 90232
      -rw-r--r--    1 root     system     46197337 May 30 07:42 Storage_Scale_<edition>-<target_version>-<architecture>-AIX-update.tgz

   2. Decompress the file.

      # gunzip Storage_Scale_<edition>-<target_version>-<architecture>-AIX-update.tgz
      # ls -ltr
      total 209128
      -rw-r--r--    1 root     system    107069440 May 30 07:42 Storage_Scale_<edition>-<target_version>-<architecture>-AIX-update.tar

   3. Extract the TAR file.

      # tar xvf Storage_Scale_<edition>-<target_version>-<architecture>-AIX-update.tar

8. Upgrade the node to a new version of IBM Storage Scale.
   1. Create the .toc file that is necessary to run the installp command.

      # pwd
      /tmp/scale_<target_version>
      # inutoc .
      # ls -ltr

   2. Install the packages of the target version of IBM Storage Scale.

      # installp -acgXd /tmp/scale_<target_version> all

   3. Verify that the target version is installed.

      # lslpp -l | grep gpfs

9. One by one, update all the nodes by following steps 6 through 8 on them.
10. Start IBM Storage Scale on the node by issuing the following command:

    # /usr/lpp/mmfs/bin/mmstartup -a

11. Check the status of the cluster.

    # /usr/lpp/mmfs/bin/mmgetstate -aL

    In the output, the GPFS status column must report active.

12. Corroborate that the target version is the one that is running on the cluster.

    # /usr/lpp/mmfs/bin/mmdiag --version

    In the output, check the value for Current GPFS build.

13. Mount any file systems that are not mounted automatically when the IBM Storage Scale daemon starts.