CES cluster setup

You can set up a Cluster Export Services (CES) environment within a GPFS cluster.

The CES shared root (cesSharedRoot) directory is needed for storing CES shared configuration data, protocol recovery, and some other protocol-specific purposes. It is part of the Cluster Export Configuration and is shared between the protocols. Every CES node requires access to the path that is configured as shared root.

To update the CES shared root directory, you must shut down the cluster, set the CES shared root directory, and start the cluster again:

mmshutdown -a
mmchconfig cesSharedRoot=shared_root_path
mmstartup -a

The recommendation for CES shared root directory is a dedicated file system. It can also reside in an existing GPFS file system. In any case, the CES shared root directory must be on GPFS and must be available when it is configured through the mmchconfig command.

To enable protocol nodes, the CES shared root directory must be defined. To enable protocol nodes, use the following command:

mmchnode --ces-enable -N Node1[,Node2...]

To disable a CES node, use the following command:

mmchnode --ces-disable -N Node1[,Node2...]

Preparing to perform service actions on the CES shared root directory file system

The CES shared root directory file system must be kept available for protocols operation to function. If a service action is to be performed on the CES shared root directory file system, perform the steps that follow.

Commands such as mmshutdown, mmstartup and mmmount, can be passed in the cesnodes node class parameter to ensure operation on all protocol nodes.

The following steps are used to perform service actions on the CES shared root file system:

Inform users of the impact to protocols. Quiesce protocol related I/O and mounts if possible. Quiesce cluster functions in progress on protocol nodes such as recalls, migrations, AFM, backup, and any policies that may be in use on the protocol nodes, or transition these cluster functions to other nodes. Finally, verify that file system quorum can be achieved by the remaining cluster nodes.
Shut down GPFS on all protocol nodes:
```
mmshutdown -N cesnodes
```
Note: Only protocol nodes need to be shut down for service of the CES shared root directory file system. However, other nodes may need to unmount the file system, depending on what service is being performed.

Protocol nodes are now ready for service actions to be performed on the CES shared root directory or the nodes themselves. To recover from a service action:

Start up GPFS on all protocol nodes:
```
mmstartup -N cesnodes
```
Make sure that the CES shared root directory file system is mounted on all protocol nodes:
```
mmmount cesSharedRoot -N cesnodes
```
Verify that all protocol services have been started:
```
mmces service list -a
```