Uploaded image for project: 'Confluence Data Center'
  1. Confluence Data Center
  2. CONFSERVER-44221

Slow user synchronisation from Crowd directory

XMLWordPrintable

      Basically, after Crowd upgrade or recreation of Crowd directory, Confluence will trigger full synchronisation which will repopulate the number of user from Crowd directory. This involved adding and updating user as per stack trace below:

      2016-09-19 21:15:31,022 INFO [scheduler_Worker-2] [atlassian.crowd.directory.DbCachingRemoteChangeOperations] getUsersToAddAndUpdate scanning [ 90786 ] users to add or update

      This operation slows down significatly due to a large user base (90140 users and 269190 of group membership)

      2016-09-19 21:16:39,947 INFO [scheduler_Worker-2] [atlassian.crowd.directory.DbCachingRemoteChangeOperations] logIfNecessary updating users - (841/90140 - 00.9%) 60032ms elapsed
2016-09-19 21:17:40,046 INFO [scheduler_Worker-2] [atlassian.crowd.directory.DbCachingRemoteChangeOperations] logIfNecessary updating users - (1287/90140 - 01.4%) 120133ms elapsed
2016-09-19 21:18:40,288 INFO [scheduler_Worker-2] [atlassian.crowd.directory.DbCachingRemoteChangeOperations] logIfNecessary updating users - (1627/90140 - 01.8%) 180375ms elapsed

      After 3 days:

      2016-09-22 14:06:29,803 INFO [scheduler_Worker-2] [atlassian.crowd.directory.DbCachingRemoteChangeOperations] logIfNecessary updating users - (52575/90140 - 58.3%) 233449890ms elapsed

      Completed after 10 days:

      INFO [scheduler_Worker-2] [atlassian.crowd.directory.DbCachingRemoteChangeOperations] logIfNecessary updating users - (90135/90140 - 100.0%) 743764777ms elapsed

      This issue does not trigger under normal circumstances (Confluence restart/reenabling Crowd directory in Confluence)

              richatkins Richard Atkins
              dazmi Dayana
              Votes:
              6 Vote for this issue
              Watchers:
              14 Start watching this issue

                Created:
                Updated:
                Resolved: