About rolling upgrade

Starting with version 2.1, NetBackup Flex Scale supports rolling upgrade of the cluster. Rolling upgrade minimizes the service and application downtime by limiting the downtime to the time it takes to stop and restart the NetBackup services on the cluster nodes.

An upgrade pre-check runs automatically at the beginning of an upgrade to determine if the system is ready for an upgrade. If the pre-upgrade check fails, you cannot proceed with the upgrade. For details about the tests performed during the pre-upgrade check, See About the pre-upgrade check.

If the upgrade fails on a node, all the nodes are automatically rolled back to the previous version. The detailed logs for rolling upgrade are located in the /log/upgrade/upgrade_3.0.log directory.

The rolling upgrade is a two-phase process. In phase A, the infrastructure packages are upgraded. The infrastructure packages are upgraded on mirror partitions on all the cluster nodes in parallel. As the upgrade happens on a copy of the original partition, the original partition is not affected and NetBackup backup and restore jobs continue running on all the nodes in the cluster. After the packages are upgraded, each node is restarted successively and the node joins the cluster with the new upgraded packages from the mirrored partition.

In phase B, the NetBackup services and containers are stopped and NetBackup and MSDP packages are upgraded on all the nodes. This phase results in downtime, which causes the ongoing backup and restore jobs to fail. After the packages are upgraded the services are restarted. You must restart the jobs once the upgrade is complete. Jobs that are configured to create checkpoints during a backup, such as file system jobs, can be suspended before an upgrade. The jobs resume after the upgrade is complete.

For details about how to perform a rolling upgrade,

For details about upgrade logs See NetBackup Flex Scale logs.