3.7.0 HP StorageWorks HP Scalable NAS File Serving Software upgrade guide HP Scalable NAS for Linux (AG513-96005, October 2009)

IMPORTANT:
Do not copy the 3.5.1 versions of the files over the 3.7.0 versions. Doing this will
cause the loss of configuration information needed by HP Scalable NAS 3.7.0.
Instead, merge the changes from the 3.5.1 files into the 3.7.0 versions of the files.
Be sure to examine the 3.5.1 files carefully to ensure that the changes are valid for
the 3.7.0 release. When modifying a 3.7.0 file, you will need to make the same
changes to the file on each server in the cluster. Also, HP Scalable NAS must be
restarted on the servers for the changes to take effect.
Verify Linux Device Mapper MPIO operations
Complete these steps to verify that Linux Device Mapper MPIO is operating:
1. Start the HP Device Mapper multipath daemon:
# /etc/init.d/multipathd start
2. Run the following command and review the output to verify that exposed LUNs
are visible on the nodes.
# multipath –ll
Replace membership partitions that are too small
In HP Scalable NAS File Serving Software 3.7.0, the minimum size for a membership
partition has been increased to 1 GB. This change was needed in part to provide
room for the new mxds datastore. If an existing membership partition is too small to
accommodate the mxds datastore, alert 13925 will be generated:
One or more membership partitions are too small to allow the mxds
data store to be created. As a result, some HP Scalable NAS File Serving
Software functionality will not be available. To correct this problem,
replace the affected membership partitions with larger partitions.
To replace a membership partition, open the Configure Cluster window and go to
the Storage Settings tab, which shows the size of each membership partition. If a
partition contains less than 1 GB, select the partition, click Replace, and then select
a new partition. If three membership partitions are configured, the replacements can
be done one-at-a-time while the cluster is online. If only one membership partition is
configured, the cluster must be offline when the partition is replaced. See “Manage
Post-upgrade steps82