Updating deployed pattern types

You can apply updates to a pattern type or a deployed pattern.

Before you begin

There are several situations where you might want to apply updates to a pattern type to a deployed pattern. For example, if you updated a pattern type or customized a plug-in that is associated with a pattern type, you might want to apply the changes to a pattern that is based on the pattern type.

You must be assigned the Workload resources administration role with permission to Manage workload resources (Full permission)

About this task

The steps in this task apply to virtual application patterns and virtual system patterns. For information about updating classic virtual system patterns, see the Related tasks section.

You can use the console or the command line interface to complete this task. For the command line information, see the Related information section.

Procedure

Click Patterns > Virtual System Instances or Patterns > Virtual Application Instances.
Select the instance that you want to update, and click Check for updates in the Pattern type section.
If an update is available, a list is displayed with the plug-ins and virtual machines that the update impacts. The type of update: rolling, in which no downtime is expected, or batch, in which downtime is expected, is listed with the update. Review the list, and then click Update or Cancel.
If you choose to update, the pattern type backs up data based on the configuration in the pattern type plug-ins, and then the update changes are applied. The update can take some time.
- If the update is successful, the Status field arrow turns green. Review the updated deployment and click Commit to update or Revert to discard the changes that were made in the update.
- If the update fails, a message is displayed in the History section: Error encountered during upgrade, automatically reverting the deployment. No action is required. The backup data is restored and the virtual application is reverted to its previous state.
Note: When you attempt a deployment upgrade for a virtual system instance, the virtual machines in that deployment restart as part of the upgrade process. Depending on various factors like the type of middleware roles that are deployed on the virtual machine, the upgrade process might take some time to complete. During this process, if any of the virtual machines take longer than 10 minutes to restart, the upgrade is attempted again. This behavior is to ensure that the upgrade request is received by all virtual machines in the virtual system instance, in case it did not happen during the initial attempt for any reason. Due to this delay, the Maestro agent might not be up and running to receive the upgrade request again, and the History section on the Virtual System Instance page shows the following message, which is just an informational message:
```
Failed while retrying deployment upgrade for deploymentId: <deployment-id>
```
This message does not denote an upgrade failure. The virtual machines might take longer to come up, depending on the type of upgrade. The upgrade process might still complete successfully, however, you can continue to commit or revert the operation according to the normal process.

Results

Your deployed instances are updated. You might notice that the previous version of the instance is shown on the Patterns > Virtual Application Instances page or the Patterns > Virtual System Instances page, depending on the pattern type that you updated. This state occurs because the appmodel.json file is not modified during the update process. Instead, your existing application model is updated to use the newer plug-ins that are provided by the pattern type. As a result, the version remains the same after the update is complete.