VERITAS Volume Manager 3.1 Administrator's Guide

Cluster Functionality
Upgrading Volume Manager Cluster Functionality
Chapter 7316
Upgrading Volume Manager Cluster
Functionality
The rolling upgrade feature allows an administrator to upgrade the
version of Volume Manager running in a cluster without shutting down
the entire cluster. To install the new version of Volume Manager running
on a cluster, the system administrator can pull out one node from the
cluster, upgrade it and then join the node back into the cluster. This is
done for each node in the cluster.
Every VxVM release, starting with Release 3.1, has a cluster protocol
version number associated with it. This is different from the release
number. The cluster protocol version is stored in the /etc/vx/volboot
file. In a new installation of VxVM, the volboot file does not exist in the
/etc/vx directory. The vxdctl init command creates this file and sets
the cluster protocol version to the highest supported version.
A new VxVM release supports two versions of cluster protocol. The lower
version corresponds to the existing VxVM release. This has a fixed set of
features and communication protocols.The higher version corresponds to
a new release of VxVM which has a new set of these features. If the new
release of VxVM does not have any functional or protocol changes, the
cluster protocol version remains unchanged, for example, in case of bug
fixes or minor changes. In this case, the vxdctl upgrade command need
not be executed.
During the rolling upgrade operation each node must be shut down and
the VxVM release with the latest cluster protocol version must be
installed. All nodes that have the new release of VxVM continue to use
the lower level version. A slave node that has the new cluster protocol
version installed tries to join the cluster. If the new cluster protocol
version is not in use on the master node, it rejects the join and provides
the current cluster protocol version to the slave node. The slave retries
the join with the cluster protocol version provided by the master node. If
the join failsat this point, the cluster protocol version on the master node
is out of range of the protocol versions supported by the joining slave. In
such a situation the system administrator must upgrade the cluster
through each intermediate release of VxVM to reach the latestsupported
cluster protocol version.
All nodes are upgraded to the latest cluster protocol version and the new
features are available.