Performing a rolling upgrade operation

During a rolling upgrade operation, the cluster remains online even though the nodes in the cluster run on different versions of PowerHA® SystemMirror® for Linux.

You can upgrade clusters from PowerHA SystemMirror Version 7.2.2 for Linux Service Pack 1, or later (source release) to PowerHA SystemMirror Version 7.2.3 for Linux, and above (target release).

Notes:

A source release is the release that is installed on the cluster nodes. A target release is the release to which you want to upgrade.
You do not need to remove the cluster when you perform a rolling upgrade operation on a cluster.

To perform a rolling upgrade operation, complete the following steps:

Download the target release to which you want to migrate. Extract the downloaded package to a local directory and change the directory to the extracted PowerHA SystemMirror directory.
Navigate to the PowerHA SystemMirror directory that contains the release to which you plan to upgrade. Run the following script to check whether the upgrade operation is allowed:
```
./installPHA --migcheck
```
Note: You must run the installPHA script of the target PowerHA SystemMirror release to which you plan to migrate.
Stop the cluster services on one node in the cluster by running the following command:
```
clmgr offline node <node> STOP_RSCT=yes manage=move
```
Notes:
- When the cluster is offline with the STOP_RSCT parameter set to yes, you cannot run clmgr command operations from that node in the cluster.
- If the manage parameter is set to move, all resources that are managed by the resource group are brought offline on the current node. The resources are brought online on the other node in the cluster.
To upgrade the node, extract the downloaded target PowerHA SystemMirror release on the node and run the following PowerHA SystemMirror installation command:
```
./installPHA -y
```
After the node is upgraded to the target release, the cluster can be brought online by running the following command:
```
clmgr online node <node>
```
Repeat steps 3 - 5 on all the nodes in the cluster.