Problem

After restarting the manager or hub, migration operations hang indefinitely at Registering or Cleaning stages. The logs show warnings: "MigrationStatus is nil for migrationId: xxx".

Root Causeh1. migrationStatuses is an in-memory map that is lost when manager restarts

After restart, Kafka replays migration status events to the handler

The handler calls SetFinished() but MigrationStatus was nil

SetFinished() silently failed, so GetFinished() always returned false

Controller kept waiting forever → timeout → rollback → fail

Solution

Add lazy initialization of migration status in the status handler by calling AddMigrationStatus() before any status updates.

Impact

- Migration operations fail after any manager restart
- Requires manual intervention to recover
- Affects production stability

Fix Details

PR: https://github.com/stolostron/multicluster-global-hub/pull/2269
Changes: 4 lines added, 1 file changed
Files Modified: manager/pkg/status/handlers/clustermigartion/managedclustermigration_handler.go