In a rolling upgrade, you upgrade software to an update release, such as from Oracle Solaris version 11.1 to version 11.2, or from Oracle Solaris Cluster version 4.1 to version 4.2 or to a version 4.1 SRU. You perform the upgrade on one node at a time. Services continue on the other nodes except for the time it takes to switch services from a node to be upgraded to a node that will remain in service.
Observe the following additional restrictions and requirements for the rolling upgrade method:
Oracle Solaris upgrade paths - You can upgrade the Oracle Solaris OS only to a new SRU or an update version of the same release. For example, you can perform a rolling upgrade from Oracle Solaris 11.1 to a later compatible Oracle Solaris 11 release, but you cannot perform a rolling upgrade from a version of Oracle Solaris 10.
Hardware configuration changes - Do not change the cluster configuration during a rolling upgrade. For example, do not add to or change the cluster interconnect or quorum devices. If you need to make such a change, do so before you start the rolling upgrade procedure or wait until after all nodes are upgraded and the cluster is committed to the new software version.
Duration of the upgrade - Limit the amount of time that you take to complete a rolling upgrade of all cluster nodes. After a node is booted into its upgraded boot environment (BE), boot the next cluster node into its upgraded BE soon as possible. You can experience performance penalties and other penalties when you run a mixed-version cluster for an extended period of time.
Software configuration changes - Avoid installing new data services or issuing any administrative configuration commands during the upgrade.
New-feature availability - Until all nodes of the cluster are successfully upgraded and the upgrade is committed, new features that are introduced by the new release might not be available.