Online migration for Engineering Workflow Management
Starting in version 5.0, when you upgrade, you can migrate data to the new version while the old production server is still online.
Determine whether to run an online migration
A fully offline migration provides easier recovery from migration errors than an online migration. If a problem occurs during a fully offline migration, the database can be returned to the premigration state by restoring the snapshot that administrators take before the migration.
Because online migration updates are made against a live server, the database cannot be restored to the current state of the data. If an unrecoverable error occurs during online migration, the database might need to be restored to a previous snapshot. As a result, any data that was created after the online migration started is lost.
To determine whether to run online migration for your repository, run the onlineMigrateEstimate command. Then, check the test results data table for guidance in interpreting the results.
Online migration process
- Determine whether to run an online migration as described in the previous section.
- To improve performance, run statistics to update the database table indexes.
- Back up your database.
- Leave the old server running.
- Install, but do not start, the new server.Note: If the new server is installed in a location that has access to the old server, the new server does not need to be configured now. If the new server is on a different computer, you must configure the Change and Configuration Management application to access the database of the old server.
- Start the online migration.
Run an online migration while the old server is running by using the onlineMigrate command on the new server and referencing the configuration files on the old server.Note: If necessary, you can stop the migration process at any time by using the onlineMigrateStop command. This command directs the online migration process to finish the current transaction and exit. Then, you have several options:
- You can perform an offline migration. The remaining work that was not performed during the online migration is completed during the offline migration.
- If you need to return the state of the database to the point where you can restart online migration from the beginning, run the onlineMigrateRevert command. This command removes partially migrated data that resulted from running the onlineMigrateStop command.
- Wait for the online migration to complete.
- Shut down the old server.
- Complete the offline migration steps.
The offline migration process can spend considerable time on the Running post addtables "com.ibm.team.scm" task without providing feedback. To obtain progress reporting for the Change and Configuration Management application before completing the offline migration, complete the following steps:
- Edit the
file by adding the following line:
<Loggers> <Logger name="com.ibm.team.scm" level="INFO"/> </Loggers>
- Run the addtables command. Then, review the log file.
When you are done obtaining progress reporting information, remove the line that you added in step a by editing the <JazzTeamServer_installation_directory>\server\conf\repotools_log4j2.xml file again.
For more information about the offline migration steps, see the Interactive upgrade guide and Upgrading IBM Engineering Lifecycle Management.
- Edit the <JazzTeamServer_installation_directory>\server\conf\repotools_log4j2.xml file by adding the following line:
- Verify the migration results.
After the online and offline migration steps are complete, you can verify the migration results by running the verify command.
- To improve performance on the new server, run statistics on your database table indexes.