Installing the upgrade using the CLI Disruptive software upgrade Recovering from software installation failures

This topic provides useful information that you will need to know when installing the upgrade.

You can use either secure copy (scp) or the SAN Volume Controller Console to copy the upgrade package to each SAN Volume Controller cluster or issue CLI commands.

If you want to use secure copy, perform the following:

  1. Once you have downloaded the software upgrade package, copy the package onto the node where the CLI is running. Issue the following to copy the package:
    scp filename admin@cluster_address:/home/admin/upgrade
    where cluster_address is your cluster IP address. You are notified of copy failures by error messages from the CLI and the SAN Volume Controller Console. If there is insufficient space on the cluster to store the software upgrade package then the copy operation will fail. If this occurs, issue the svctask cleardumps command to make space for the upgrade package, then repeat the copy operation.
  2. After a successful copy of the file, issue the svcservicetask applysoftware -file filename command, where filename is the name of the file that you copied the software upgrade package too. This command starts the installation of the code. The installation process will fail if a node is not present and if the node is not paired with another node in an I/O group. You can, however, use the -force option to override this restriction if you are prepared to lose access to data during the upgrade.
    Note:Note: The installation process will only fail when some paths between the host systems and the cluster are not available. Data access can be lost temporarily during the upgrading process. You can prevent this if, before you start the installation, you issue a datapath query device on each host system to ensure that all paths are available. See the IBM TotalStorage Subsystem Device Driver: User's Guide for more information about datapath query commands.
    Attention: Attention: The order in which the nodes are upgraded depends on the following:
  3. To verify that the upgrade was successful, you can perform any one of the following steps:
    • The code level is distributed to all the nodes that are in the cluster. The nodes, in turn, are then restarted. If all the nodes successfully restart with the new code level, the new version is committed and the cluster vital product data (VPD) is updated to new level of code.
    • The software upgrade is complete when the cluster verifies the upgrade package. To determine whether the upgrade has completed, you must either display the software version in the cluster VPD, or look for the Software upgrade complete event in the SAN Volume Controller error or event log. If the node does not restart automatically during the upgrade, you should repair or manually delete that node from the cluster to complete the backout process.
    • Alternatively, you can also either perform the following steps:
      1. Issue the svctask dumperrlog command to dump the contents of the error log to a text file. You can also use this command to delete unwanted error log dumps from the cluster.
      2. Once you have the contents of the error log dumped into a text file, verify that there were no errors in the text file. If there are no errors, you have successfully upgraded the software and output similar to the following is displayed in the log file:
        Upgrade completed successfully
      3. Issue the svcinfo lsnodevpd command for each node. You should see that the software version field has been updated.
Related topics Disruptive software upgrade Recovering from software installation failures

(C) Copyright International Business Machines Corporation 2003, 2004