Details
-
Bug
-
Resolution: Support Request
-
Medium
-
None
-
5.8.6, 5.8.9
-
None
Description
We just migrated from Confluence 5.5.6 to Confluence 5.8.6. Unfortunately, our Microsoft LDAP (lazy loading) sync is not ending, after we defined the 'objectGUID' as Unique User Attribute on LDAP config. It is now migrating users for 5 days and 21 hours and does not seem to come to an end (current status 58% of 142,000 users). The amount of users migrated for an interval of one minute is shrinking from hundreds to only 5 per minute.
What we observed on system backend:
- Two CPU cores are used for 100%.
- According to an Oracle AWR report, most frequent statement (CPU times 74s) is 'select * from ( select journalent0_.entry_id as entry_id journalent0_.journal_name as journal_2_ journalent0_.creationdate as creation3_ journalent0_.type as type journalent0_.message as message from journalentry journalent0_ where (journalent0_.entry_id>:1 )and(journalent0_.journal_name=:2 )and(journalent0_.creationdate<=:3 ) order by journalent0_.entry_id asc ) where rownum <= :4*
- According to Performance Profiling, slowest / most time consuming method is '2015-08-26 09:11:19,765 DEBUG [scheduler_Worker-5] [atlassian.util.profiling.UtilTimerStack] log [1848ms] - PersonalInformationManager.getOrCreatePersonalInformation()'
Why is this method performing so slow?
Please find attached an excerpt of the page profiling.
For us this is a very serious issue since we really need a full working LDAP integration for user and group management!
The same issue is also valid for Confluence 5.8.9, which we also tested.
Attachments
Issue Links
- is related to
-
CSP-155974 Loading...