PRIMECLUSTER Enterprise Edition 4.3A30B Installation Guide

B.2.1 Upgrading from PRIMECLUSTER Enterprise Edition 4.3A20

1. Boot all the cluster management servers in single user mode.

   # shutdown now <Return>

2. When update the basic software from existing environment, Updating basic software, see the following document Operating Update manual.

3. Proceed the following steps on each cluster management server.

   1. Create a backup directory.

      # mkdir /<mydir> <Return>

   2. Back up the PRIMECLUSTER operating environment.

      # cp -p /opt/FJSVwvbs/etc/Plugin.html /<mydir> <Return>

      In case of Red Hat Enterprise Linux 5 (for Intel64) or
      Red Hat Enterprise Linux 6 (for Intel64) ,
      back up the PRIMECLUSTER operating environment.

      # cp -p /opt/FJSVwvbs/etc/webview.cnf /<mydir> <Return>
      # cp -p /opt/FJSVwvbs/etc/.policy /<mydir> <Return>
      # cp -p /opt/FJSVwvbs/etc/wvlocal.cnf /<mydir> <Return>

   3. Insert CD and mount the CD-ROM device.

      # mount /media/cdrom <Return>

      <CDROM_DIR> will be used as the mount point.

   4. Execute the following script then delete part of the PRIMECLUSTER package.

      # cd <CDROM_DIR>/Tool <Return>
      # ./upgrade_uninstall <Return>
      Are you sure to remove a part of PRIMECLUSTER from your system (y or n) ? y <Return>
      :
      :
      The uninstallation finished successfully.

   5. Execute the following cluster_install script, and install the package or override it.

      # cd <CDROM_DIR>/Tool <Return>
      # ./cluster_install -e EE-M <Return>
      :
      :
      The installation finished successfully.

      Note

      • The following message might be output:

        # ./cluster_install -x xx <Return>
        INFO: no package to update

        This message indicates that the newer version of all the packages is installed, so it is not necessary to upgrade.

      • While executing the cluster_install script, the following message might be output

        # ./cluster_install -x xx <Return>
        Installing package <XXXXXXXXXXX> ... skipped.

        This message indicates that the same version of all the packages is installed, so it is not necessary to upgrade.

   6. Eject CD.

      # cd / <Return>
      # eject cdrom <Return>

   7. Restore the PRIMECLUSTER operating environment that was backed up at step 2.

      # sed s/"[ ]*WIDTH.*HEIGHT.*ALIGN.*"/"  WIDTH = \"810\" HEIGHT = \"430\" ALIGN = \"TOP\""/g 
         /<mydir>/Plugin.html > /opt/FJSVwvbs/etc/Plugin.html  <Return>

      In case of Red Hat Enterprise Linux 5 (for Intel64) or
      Red Hat Enterprise Linux 6 (for Intel64) ,
      restore the PRIMECLUSTER operating environment that was backed up at step 2.

      # cp -p /<mydir>/webview.cnf /opt/FJSVwvbs/etc/webview.cnf <Return>
      # cp -p /<mydir>/.policy /opt/FJSVwvbs/etc/.policy <Return>
      # cp -p /<mydir>/wvlocal.cnf /opt/FJSVwvbs/etc/wvlocal.cnf <Return>

4. After completing step 3 on all the cluster management servers, reboot all the cluster management servers.

   # shutdown -r now <Return>