This article has been archived and is no longer maintained by ForgeRock.
The following error is shown in the Errors log when the replication server fails to start:[22/Oct/2016:16:29:52 +1000] category=SYNC severity=ERROR msgID=org.opends.messages.replication.274 msg=The following log '/path/to/opendj/changelogDb/2.dom/10566.server' must be released but it is not referenced." [22/Oct/2016:16:29:52 +1000] category=SYNC severity=ERROR msgID=org.opends.messages.replication.11 msg=The replication server failed to start because the database /path/to/opendj/changelogDb could not be read : Could not get or create replica DB for baseDN 'dc=example,dc=com', serverId '10566', generationId '72390'
You may also experience issues with OpenAM when this happens since OpenDJ will not be able to save any configuration changes you make. You will see generic "Unwilling to Perform" errors such as the following in the OpenAM Configuration log:amSMSEmbeddedLdap:22/10/2016 16:31:56:289 PM GMT: Thread[http-nio-0.0.0.0-9443-exec-16,5,main]: TransactionId[daaeffd3-d4a3-4324-8f1d-865e43f34773-168] ERROR: SMSEmbeddedLdapObject.modify: Error modifying entry ou=AgentUsers,ou=default,ou=OrganizationConfig,ou=1.0,ou=sunidentityrepositoryservice,ou=services,o=agentusers,ou=services,dc=openam,dc=example,dc=com by Principal: id=amadmin,ou=user,dc=openam,dc=example,dc=com, error code = Unwilling to Perform
If this happens, you will need to check the OpenDJ logs to verify it is the same issue.
When the changelogDb file size is a multiple of 256 (the internal block size), the server fails to compute the newest record after being restarted.
This issue can be resolved by upgrading to OpenDJ 3.5 or later; you can download this from BackStage.
You can follow this process as a workaround until your next restart:
- Stop the OpenDJ instance.
- Move the changelogDB directory to a temporary location to allow a new one to be re-created.
- Restart the OpenDJ instance and wait for the changelogDB directory to be rebuilt.
Once the new changelogDB directory has been rebuilt and you have verified replication is working as expected, you can remove the old changelogDB directory from the temporary location.