2 Upgrading an Existing ACSLS HA 8.4 Solaris Environment to Use ACSLS 8.5.1

Use the following procedure to upgrade an existing ACSLS HA 8.4 Solaris environment to use Solaris 11.4 SRU14, ACSLS 8.5.1, and Solaris Cluster 4.4.
Review all steps before performing the procedure and ensure that you perform all steps in the order listed.

Note:

  • Be aware of component versions as updates proceed. Several reboots and additional steps for Solaris are required.
  • Do not uninstall ACSLSHA 8.4 at any time during this procedure.
  • Reference the following publications:
    • ACSLS 8.4 Installation Guide
    • ACSLS 8.5 Installation Guide
    • ACSLS 8.4 HA Cluster Installation, Configuration, and Operation
  1. Quiesce all activity at any client nodes that may have access to ACSLS.
  2. Shut down ACSLS on the primary node:
    # su – acsss $acsss shutdown
  3. Set the Solaris Cluster auto start feature to false to prevent the cluster from attempting start in the event of a reboot:
    # clrg set –p Auto_start_on_new_cluster=false
  4. Suspend ACSLS cluster group resources and ensure that ACSLS is shut down:
    # clrg suspend acsls-rg
    # su – acsss $acsss shutdown
  5. On the node currently mounted to the zpool, perform either a backup of ACSLS 8.4 or an export of the database if you chose to restore or import your library configuration to ACSLS 8.5.1.
    Optionally, you can skip this step and simply perform an audit after executing acsss_config with ACSLS 8.5.1.

    Note:

    You are only required to back up ACSLS from one node.
  6. Boot both nodes with the –x option to ensure that the cluster software is not started:
    # boot –x
  7. Uninstall the cluster software from both nodes:
    # scuninstall –r –v
  8. Reboot both nodes.
  9. Update the first node:
    1. Ensure that the zpool acslspool remains imported and that /export/home is mounted on the node.
    2. Uninstall ACSLS 8.4:
      # /opt/ACSLS/pkg_uninstall

      Note:

      Do not uninstall ACSLSHA 8.4 at any time during this process.
    3. Update Solaris to 11.2 SRU15 (latest)
    4. Update Solaris to 11.3 Base (reboot required)
    5. Update Solaris to 11.3 SRU36 (reboot required)
    6. Update Solaris to 11.4. Base (reboot required and a new boot environment is created).

      Note:

      The Solaris 11.4 update will fail if you have not successfully uninstalled ACSLS 8.4.
    7. Update Solaris to 11.4 SRU14 (reboot required).
  10. Update the second node. Repeat all instructions in the previous step for this node.
  11. Reboot both nodes. They are now running Solaris 11.4 SRU14.
  12. Install ACSLS 8.5.1 and Solaris Cluster 4.4 on the first node:
    1. Utilize the ACSLS 8.5.1 Installation Guide to install ACSLS 8.5.1 and either perform a restore from the backup that you created in Step 5 or configure your library (acsss_config).
    2. Install Solaris Cluster 4.4 using the cluster installation instructions provided in the ACSLSHA 8.4 Installation Guide, but do not run the scinstall utility at this time. You will run scinstall from the second node in step 13.
    3. Export the zpool acslspool (/export/home is unmounted).
  13. Install ACSLS 8.5.1 and Solaris Cluster 4.4 on the second node:
    1. Import the zpool acslspool (/export/home is mounted).
    2. Utilize the ACSLS 8.5.1 Installation Guide to install ACSLS 8.5.1.
      If you performed a restore of ACSLS on the first node (in step 12) do not run acsss_config or restore on the second node. The second node will import the database from the common /export/home mount point when it is started as the primary or when fail over occurs.
    3. Install Solaris Cluster 4.4 using the cluster installation instructions provided in the ACSLSHA 8.4 Installation Guide and run the scinstall utility, specifying the same interconnect network devices previously used with Solaris Cluster 4.2.
    4. Run the ACSLSHA command acsAgt configure (described in the ACSLSHA 8.4 Installation Guide) on the node on which the zpool acslspool is imported.
      Respond to the prompts for network devices in the same manner in which they were assigned for the original ACSLSHA 8.4 environment.
  14. Reboot both nodes and ensure that the cluster resources acsls-rs, acsls-storage, and logicalHost are online to one host using the clrs status command.