-
Bug
-
Resolution: Fixed
-
Medium
-
9.3.0, 9.4.0, 9.3.1, 9.3.2, 9.5.0, 9.4.1, 9.4.2, 9.4.3
-
1
-
Severity 3 - Minor
-
1
-
Issue Summary
Starting with Bitbucket 9.3, when managing a cluster, you may encounter issues if you attempt to manually lock the cluster or perform backups while one of the nodes is shut down or in a stopped state. In such cases, the lock might not be acquired successfully. Even if the lock is acquired, you may face difficulties unlocking the instance later, which requires a instance restart.
Steps to Reproduce:
- Setup a Bitbucket Data Center Cluster in version 9.3 or higher.
- Shutdown one of your nodes.
- Manually lock your instance with below endpoint
curl -s \ -u ${BITBUCKET_BACKUP_USER}:${BITBUCKET_BACKUP_PASS} \ -X POST \ -H "Content-type: application/json" \ "${BITBUCKET_URL}/mvc/maintenance/lock"
- Check logs and see if lock was able to be acquired.
- Try to unlock you instance with no success if lock was acquired.
Expected Results:
Lock is able to be acquired and instance unlock successful.
Actual Results:
Below error is thrown in the logs when lock is ran for the first time and when trying to unlock using the token found in the logs:
2025-02-07 03:00:07,720 DEBUG [http-nio-7990-exec-4] admin @RMTUB2x799x390x0 10.224.217.110,10.224.217.84 "POST /mvc/maintenance/lock HTTP/1.0" c.a.s.i.a.DefaultAuthenticationService username_test: Authenticator 'com.atlassian.bitbucket.server.bitbucket-authentication:crowdHttpAuthHandler' authenticated admin (1) 2025-02-13 13:19:01,735 ERROR [http-nio-7990-exec-7] admin @RMTUB2x799x390x0 10.224.217.110,10.224.217.84 "POST /mvc/maintenance/lock HTTP/1.1" c.a.s.i.w.HandlerExceptionResolverImpl ExecutorService[bitbucket.core] is shutdown! In order to create a new ExecutorService with name 'bitbucket.core', you need to destroy current ExecutorService first! at com.atlassian.stash.internal.maintenance.ClusterMaintenanceLock.lock(ClusterMaintenanceLock.java:56) at com.atlassian.stash.internal.maintenance.DefaultMaintenanceService.lock(DefaultMaintenanceService.java:172) at jdk.proxy3/jdk.proxy3.$Proxy329.lock(Unknown Source) at com.atlassian.stash.internal.web.maintenance.MaintenanceController.lock(MaintenanceController.java:185)
Workaround:
Restart instance and run locks with all nodes up.
| Form Name | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|