Group-level data migration
Data migration is an advanced feature with the potential for data loss and/or data corruption. This feature must be leveraged with caution, with proper understanding and planning.
The group-level data migration feature allows you to move message data associated with a specific group to a new repository (a different database instance). This may be useful when needing to scale your existing data / databases horizontally. It allows you to split a large repository database.
In order to keep the data payload as small as possible (and limit the performance impacts of the migration), current functionality limits the data being moved to:
All
queued
messagesAll
delayed
messagesThe previous 2 days of all other queue types
To start a migration, select the groups
UI within Connexion, select the group you wish to migration, and click the Migrate Group...
link.
This will display the migration dialog, where you can select the target repository. Click the Run Migration
button to start moving data. Any running channels will be automatically stopped and locked prior to starting the migration.
Note that this dialog retains a history of previous migrations (which is cleared when the Connexion service restarts). The top section of the dialog displays a grid of current and previous migrations, while the bottom section displays the output from the selected migration.
The Migrate Group
dialog will automatically poll for migration status and display updates in the lower pane of the dialog.
Once the migration completes, a message box is displayed and the status updates to complete
.
Any previously running channels will be restarted, and channel locks cleared.
The message data left behind (if any) will be purged when database maintenance is next run.