This section outlines the procedure to be followed to upgrade to the MS SQL Failover Cluster Instance Enterprise Workloads agent from agent version 6.x.
Before you begin: Ensure you have administrative privileges on all nodes of the failover cluster.
1. Remove the Agent Service from the Cluster Role
First, you must remove the existing agent's service resource from the cluster role itself.
Open Failover Cluster Manager on your server.
In the left-hand navigation pane, select Roles.
Find and select your target SQL Server Role from the central pane.
In the Resources tab at the bottom of the window, find the resource named Hybrid workloads agent client service.
Right-click the Hybrid workloads agent client service and select Remove from the context menu. Confirm the action if prompted.
2. Uninstall the Agent from the Active Node
Next, you must uninstall the main agent program from the node that is currently hosting the SQL Server role (the active node).
Identify the owner of the SQL Server Role in Failover Cluster Manager.
Log in directly to that active node.
Open Control Panel and navigate to Programs and Features (or open Settings > Apps).
Find the Hybrid workloads agent in the list of installed applications.
Select the agent and click Uninstall.
Follow the prompts in the uninstallation wizard to complete the removal.
3. Delete the Agent Service from Passive Nodes
The uninstaller only runs on the owner node. You must now manually delete the service from all other (passive) nodes in the cluster.
Log in to each passive node in your cluster.
Open Command Prompt (cmd.exe) or PowerShell as an Administrator.
Type the following command and press Enter:
sc.exe delete PhoenixCPHService
Verify that you see a [SC] DeleteService SUCCESS message.
Repeat this process for every passive node.
4. Clean Up Junction Point from All Nodes
A final cleanup step is required to remove the data directory junction point. You must perform this step on all nodes in the cluster (both the active node and all passive nodes).
While logged in as an Administrator on each node, open Command Prompt (cmd.exe) or PowerShell.
Run the following command to delete the junction point:
rmdir C:\ProgramData\Phoenix
Confirm the action if prompted. Ensure this directory is removed from all nodes.
5. Re-register the FCI
After you have completed all the cleanup steps above on all cluster nodes, you are ready to proceed with the new agent.
Re-register your FCI server with the new EnterpriseWorkloads Agent.
Important: You must perform the following procedure every time you upgrade the MS SQL FCI Enterprise Workloads Agent from version 7.x. The upgrade process does not automatically bring the cluster resource online, so you must bring it online manually using Failover Cluster Manager.
Procedure
Open Failover Cluster Manager on your server.
In the left-hand navigation pane, select Roles.
From the main window, select the specific SQL FCI Role for the agent you just upgraded.
At the bottom of the window, select the Resources tab.
Under the "Other Resources" group, find the service named Druva-Enterprise Workload Service.
Right-click the Druva-Enterprise Workload Service and select Bring Online from the context menu.
Note: Make sure that in the Failover Cluster Manager, the owner node of the MSSQL Server is the same node from which the upgrade was triggered. If it is not, then stop the Druva-Enterprise Workload Service from services and start it on the new owner node. Then, bring the Druva-Enterprise Workload Service online.