Removing a managed node profile from a deployment environment

Edit online

To remove a managed node profile from a IBM® Business Automation Workflow deployment environment, first remove the node from the environment and then delete the profile. No database updates are required.

Before you begin

For more information about the manageprofiles command-line utility, see manageprofiles command-line utility.

Note: The deployment environment wizard does not support modification to the deployment environment. If you attempt to modify the deployment environment with the deployment environment wizard, the cluster members become partially nonfunctional. Instead, you can use the BPMConfig command to modify your deployment environment. For more information, see BPMConfig command-line utility.

Procedure

Stop the deployment manager and all the running server processes in the managed node, including the node agent process.
On the deployment manager, run the following command to remove the cluster members from the Business Automation Workflow deployment environment:
```
BPMConfig -delete -profile dmgr_profile_name -de de_name -node node_name
```
Running this command does not delete the node. It helps ensure that REST endpoints and JMS connection factory provider endpoints that refer to the cluster members are updated. For more information about the command, see BPMConfig -delete.
Restart the deployment manager.
On the managed node machine, run the following command:
```
manageprofiles -delete -profileName node_profile_name
```
By default, a prompt for the credentials to connect to the running deployment manager is displayed. This prompt times out after 1 minute and there are no manageprofiles parameters to specify the credentials on the command line. To avoid the prompt, set the following properties in node_profile/properties/soap.client.props:
- com.ibm.SOAP.loginUserid=admin_user
- com.ibm.SOAP.loginPassword=admin_password
Running this command deletes all servers that are not Business Automation Workflow servers, including web servers, and all application module mappings that refer to those servers. It also deletes the node itself from the WebSphere® Application Server cell.
Optional: If the removed node hosted a web server, you must update the web server plugin.
For instructions, see the Plug-ins configuration documentation.
Optional: Manually remove any virtual host aliases that are no longer needed.
In the administrative console, go to Environment > Virtual hosts > default_host > Host Aliases and delete aliases that you no longer need.
Optional: If you have errors, complete the managed node removal by running the cleanupNode command and removing references from the registry.
- If the manageprofiles command fails to connect to the deployment manager, the following message is displayed: INSTCONFPARTIALSUCCESS: The profile no longer exists, but errors occurred. The node_profile/logs/_nodeuninst.log file contains the following error: ADMU0040E: Exception on MBean invocation WebSphere:name=AdminOperations,process=dmgr.
- If the deployment manager was not running at all, the following message is displayed: INSTCONFSUCCESS: Success: The profile no longer exists. The node_profile/logs/_nodeuninst.log file contains the following error: ADMU0006E: Exception creating Deployment Manager connection.
1. Confirm that the deployment manager is running, then run the following command on the deployment manager to complete the managed node removal:
```
cleanupNode node_name -profileName dmgr_profile_name -username admin_user -password admin_password
```
  For more information, see cleanupNode command.
2. On the computer where the node was deleted, run the following command to remove references in the registry to deleted profiles:
  - ./manageprofiles.sh -validateAndUpdateRegistry
  - manageprofiles.bat -validateAndUpdateRegistry
Remove the managed node profile directory manually by using operating system commands.
The manageprofiles command does not entirely remove the directory because it still contains log files.
Clean up the no longer existing node from the Event Manager > Monitor panel by deleting the corresponding line from the LSW_EM_INSTANCE (at BPMDB) table.
1. Login to Process Admin Console and select Event Manager > Monitor. Look at the Scheduler ID column and note the name of the Scheduler that is associated with the deleted node.
2. At Database layer, back up the LSW_EM_INSTANCE table.
3. Delete the row with the following SQL statement:
```
db2 "DELETE <schema>.LSW_EM_INSTANCE WHERE NAME='<Scheduler_name>'
```
For more information, see Understanding and Tuning the Event Manager.

Note: All schedulers that have ever run, are listed in lsw_em_instance and are shown on the Event Manager monitor page. If a scheduler is disabled, it updates or inserts its row at start, so that Event Manager monitor users recognize that the machine is there, but the scheduler is disabled. The difference between disconnected, disabled, and paused is only visible by moving your mouse over the red light that is associated with the scheduler instance on the Event Manager monitor page. If a particular machine no longer exists, and you do not want to see it on the Event Manager monitor page anymore, you can carefully remove its row from the lsw_em_instance table. The row is recreated when a machine with that name starts up again.