Preparation and prerequisites

  1. Before the online upgrade, make sure that all IBM Storage Scale System nodes are active by running the following command from one of the cluster nodes:
    /usr/lpp/mmfs/bin/mmgetstate -N $(/usr/lpp/mmfs/bin/mmlsnodeclass -Y | 
/usr/lpp/mmfs/bin/mmyfields nodeClass allMembers | 
grep -iE "ems|ess3500_x86_64|ess3200_x86_64|ess3500_x86_64|ess5k_ppc64le|s6k_x86_64" | 
awk '{ if (NF > 1) print $(NF-1)}' | paste -d',' -s)
  2. Download the target IBM Storage Scale System release from IBM Fix Central.
  3. If IBM Storage Scale native REST API is enabled on the IBM Storage Scale System cluster, disable the SSH/SCP wrapper during the upgrade. To return to SSH and SCP, you can issue the following command:
    mmchcluster -r /usr/bin/ssh -R /usr/bin/scp
  4. Copy the IBM Storage Scale System release image, which is downloaded from IBM Fix Central, to the /home/deploy directory on the management server.
Remember:
  • The GUI is stopped during the management server upgrade.
  • Firmware is not automatically updated. Consider this while you plan the maintenance window in the environment to be upgraded.
Important:
  • An update must be done over the management network. Use the correct hostname.
  • High-speed network might become inaccessible during the upgrade process.
Note: Offline upgrade assumptions (management server or protocol nodes only)
  • You assume the risks of potential quorum loss.
  • The IBM Storage Scale GUI and collector must be down.