HVM Deployment Upgrade Guide¶

This guide provides a cohesive, end-to-end upgrade procedure for a complete HVM deployment. An HVM deployment consists of the HPE Morpheus Software Manager (appliance), one or more HVM clusters, the HPE Morpheus Software Agents on cluster hosts, and host-level packages. Each component has its own upgrade mechanism, but they must be performed in the correct order.

Upgrade Order¶

The recommended upgrade sequence is:

Step 1: Upgrade the Morpheus Manager¶

The HPE Morpheus Software Manager appliance is upgraded first. HVM clusters operate autonomously during this window — quorum, heartbeat monitoring, and VM failover continue regardless of Manager availability.

Procedure:

1. Back up the Manager database and configuration before starting. See Backing Up and Restoring HPE Morpheus Software Appliance.

2. Follow the appropriate appliance upgrade procedure for your topology:

   • Single node: Single Node/AIO Appliance Upgrades

   • 3-node HA: 3-Node HA Upgrade

   • Full HA: Full HA Upgrade

3. After the Manager is upgraded and accessible, verify:

   • The Manager UI is reachable and login succeeds

   • Navigate to Infrastructure > Clusters and confirm all clusters are visible

   • Agent connectivity: all hosts show recent lastAgentUpdate timestamps

Step 2: Rolling Cluster Update¶

Once the Manager is running the new version, trigger the cluster update. This is an orchestrated operation that processes each host sequentially:

What the rolling update does per host:

1. Upgrades the Morpheus Agent (if the host agent is below the required minAgentVersion for the update)

2. Enters maintenance mode — Live-migrates all VMs off the host

3. Executes update scripts — Applies host package updates, configuration changes, and layout-specific operations

4. Exits maintenance mode — Host returns to service and can receive VMs again

5. Proceeds to next host

After all hosts complete, any postUpdate scripts run across the cluster.

Triggering the update:

1. Navigate to Infrastructure > Clusters > [Your Cluster]

2. If an update is available, the UI will indicate it. Select the update action.

3. Monitor progress in the cluster history/activity feed

Prerequisites before starting:

Failure handling:

If the update fails on any host:

• Rollback scripts execute on the failed host and all previously updated hosts

• The cluster returns to its pre-update state

• Review the error, resolve the underlying issue, and retry

For full details on rolling updates, failure handling, and rollback, see Upgrading Clusters.

Layout Upgrades (1.2 → 1.3)¶

If upgrading from layout 1.2 to 1.3, the rolling update additionally:

• Disables and removes Pacemaker services

• Activates the HPE Morpheus Software Agent QuorumCheckService

• Distributes quorum configuration to all agents

After this upgrade, pcs commands are no longer applicable. Use corosync-quorumtool and the agent /quorum endpoint for cluster operations.

Agent Upgrades¶

Agent upgrades are handled automatically during the rolling cluster update when needed. Manual agent upgrades are typically not required.

If you need to verify agent versions outside of a cluster update:

• Navigate to Infrastructure > Clusters > [Cluster] > Hosts

• Check the agent version reported for each host

• The cluster update will upgrade any agent below minAgentVersion before applying host updates

Host Package Updates¶

Host-level package updates (kernel, libraries, HVM OS packages) are applied by the update scripts during the rolling update. These are not arbitrary apt upgrade operations — they are curated package sets defined by the cluster layout and delivered through the update mechanism.

Step 3: Post-Upgrade Verification¶

After both the Manager and cluster updates complete, verify the full stack:

Manager verification:

• UI accessible, all expected clouds and clusters visible

• Check Administration > Health for any service warnings

Cluster verification:

1. Quorum health:

   curl -k https://localhost:7443/quorum
   

   Confirm quorum is ACHIEVED with all nodes ONLINE.

2. Corosync membership:

   corosync-quorumtool -s
   

   Confirm Quorate: Yes and all expected nodes listed.

3. DLM status:

   dlm_tool status -v
   

   Confirm all nodes show member=1.

4. Datastore mounts:

   mount | grep gfs2
   

   Confirm all HPE Clustered Datastores are mounted.

5. UI Quorum panel (Infrastructure > Clusters > [Cluster] > Summary > Quorum):

   • Quorum: ACHIEVED

   • All nodes: ONLINE

   • Coordinator: assigned

   • Lockspaces: OK

VM verification:

• All VMs running and accessible

• No unexpected VM migrations or restarts in the activity log

Troubleshooting¶

Related Documentation¶

• Upgrading — Manager appliance upgrade procedures

• Upgrading Clusters — Detailed cluster update mechanics and layout transitions

• Host Maintenance & Evacuation — Host maintenance mode operations

• Troubleshooting & Diagnostics — Cluster diagnostic commands

• Backing Up and Restoring HPE Morpheus Software Appliance — Backup and restore procedures