You can migrate your WebSphere® Extended
Deployment Compute Grid or Feature Pack for Modern Batch servers to WebSphere Application Server Version 8.5.
Before you begin
- To Migrate WebSphere Extended
Deployment Compute Grid Version 6.1.1.3 or above to Version 8.5, you must have WebSphere Extended Deployment
Compute Grid Fixpack 3 and cumulative interim fix PM39203 installed.
- You must first migrate the deployment manager before migrating
application server nodes.
About this task
Migrate the schedulers one at a time and then migrate
the endpoints. Repeat the following steps for each Compute Grid or batch server:
Procedure
- Stop the servers.
- Install WebSphere Application Server Version 8.5.
- Migrate your old application server nodes to WebSphere
Application Server Version 8.5. You can use the migration tools to migrate your old deployment
manager to Version 8.5.
- Start the node agent and wait for it to synchronize.
- Synchronize and start the Job Scheduler servers.
Avoid trouble: Do not start endpoint servers.
gotcha
Log into your
administrative console and ensure the newly migrated node is properly
synchronized with the deployment manager. Start the Job Scheduler
servers on the node.
- Run the restore command for the scheduler. For
example:
Change directories to your scheduler node profile bin dir,
and run restore for the scheduler
wsadmin -lang jython -f migrateConfigTo85.py --restore -nameOfProfile <profile>
-configBackupDir <PathToBackupLocation> -node <scheduler_node_name>
- Stop all other Job Schedulers on all other nodes. You
must stop all non-migrated Job Schedulers.
Note: This action will put
your configuration into a state where it is no longer highly available.
If you are approaching a scheduled critical batch window, consider
backing up your work and restoring the production environment.
- Start the migrated endpoint server. When the
migrated job scheduler server is the only job scheduler running, you
can start the migrated endpoint server. It will co-exist with all
other currently running endpoint servers.
What to do next
Migrate the remaining nodes. For each remaining node:
- Stop all servers running on that node.
- Start the node agent and restore the batch configuration.
- Synchronize the node with the deployment manager.
- Start any job scheduler and end point cluster members on the node.