Unsupported upgrade scenarios with the viosupgrade tool

Learn about the scenarios that are not supported for upgrade with the Virtual I/O Server (VIOS) upgrade tool.

Full cluster restore in a single instance is not supported

The VIOS upgrade tool supports backup and restore of VIOS at cluster level with the -c flag. However, full cluster restore in a single instance is not supported in the VIOS 3.1 release. Hence, when you upgrade VIOS nodes that belong to a Shared Storage Pool (SSP) cluster, irrespective of the number of nodes, you must upgrade a few nodes at a time, while keeping the cluster running on the other nodes. Failure to do so might result in the loss of cluster connectivity.

For example, for a 4-node cluster, you can upgrade 1, 2, or 3 nodes, while keeping at least one node active in the cluster. After the successful installation of the first set of nodes, you can choose to upgrade the second set of nodes in the cluster.

Note: For single node clusters, it is not possible to use the VIOS upgrade tool to upgrade and restore the cluster. You must manage single node clusters manually. Alternatively, you can add one or more nodes to the SSP cluster before you initiate the upgrade process on the first node.

Rootvg LV backed vSCSI disk backup restore not supported

Currently, the viosbr command does not support virtual SCSI (vSCSI) disks that are created on the rootvg disks of the Virtual I/O Server. Hence, the VIOS upgrade tool cannot be used to restore vSCSI mappings if Logical Volumes (LVs) are created from a rootvg disk. You must move the vSCSI disks from the rootvg to other volume groups before you initiate the upgrade process. Alternatively, you can start the installation on an alternative disk by preserving the current rootvg.

For more information about using Logical Volume Manager (LVM) commands (such as the cplv command) to migrate these vSCSI logical volumes, see IBM Moving a JFS/JFS2 File System to a New Volume Group.