Details
-
Bug
-
Resolution: Unresolved
-
Medium
-
None
-
5.5.6
-
10
-
Severity 2 - Major
-
1
-
Description
Bug Description
Due to multiple reason explain here Confluence set the value for attribute_name "com.atlassian.crowd.directory.sync.issynchronising" to "true" which made it think that it is still synchronising against its user directories even after restarting the instance.
This cause Confluence not to run the initial synchronisation during the start-up process and it won't be able to synchronise automatically afterwards.
Finding
In some cases, if Confluence synchronisation was incomplete users from the affected directory won't be able to log-in as the users are not synchronize yet.
If this occurred to a Confluence instance that use Crowd SSO, user won't be able to log-in using its internal admin users to manually synchronise the Crowd directory.
Recommendation
Confluence should check the value of "com.atlassian.crowd.directory.sync.issynchronising" during restart process and set it to "false" automatically if it is set to "true" to resolve this issue
Workaround
- Shutdown Confluence
- Execute the following SQL query to the database and take note of the problematic directory ID
SELECT * FROM cwd_directory;
- Execute the following SQL query to ensure that Confluence won't run the synchronization after restart
UPDATE cwd_directory_attribute SET attribute_value = 'false' WHERE attribute_name='com.atlassian.crowd.directory.sync.issynchronising' AND directory_id = <from the previous query>;
- Restart Confluence
Attachments
Issue Links
- is related to
-
CONFSERVER-25426 Maximum time of synchronization implementation
- Closed