The initial configuration and import of the LDAP into confluence ran really fast (~2 hours). We had to update the field mapping to remove the displayName (our ldap displayName field does not contain a readable name). To do this we entered a field name that does not exist in the LDAP (ex: "User Display Name Attribute:*" = "dummyValueThatDoesNotMatchLDAP"). After making this change the LDAP starting a full sync and ran well for a number of hours. However it has been slowing down ever since. Currently we are processing roughly 4 users a minute. At this rate it will be another 4 days until the sync finished (We started this 4 days ago). Also the processing rate is slowing so I'm concerned it will not finish. Our CPU usage is now almost maxed out.
I'm seeing this in the log files:
2016-03-07 16:19:46,738 INFO [scheduler_Worker-1] [atlassian.crowd.directory.DbCachingRemoteChangeOperations] logIfNecessary updating users - (65097/104641 - 62.2%) 338414832ms elapsed
2016-03-07 16:20:21,482 WARN [scheduler_Worker-1] [confluence.util.profiling.DurationThresholdWarningTimingHelperFactory] logMessage Execution time for publishing event com.atlassian.crowd.event.user.UserEditedEvent@c213059 took 9137 ms (warning threshold is 5000 ms)
2016-03-07 16:20:38,389 WARN [scheduler_Worker-1] [confluence.util.profiling.DurationThresholdWarningTimingHelperFactory] logMessage Execution time for publishing event com.atlassian.crowd.event.user.UserEditedEvent@c213059 took 6994 ms (warning threshold is 5000 ms)
2016-03-07 16:20:58,914 WARN [scheduler_Worker-1] [confluence.util.profiling.DurationThresholdWarningTimingHelperFactory] logMessage Execution time for publishing event com.atlassian.crowd.event.user.UserEditedEvent@c213059 took 8889 ms (warning threshold is 5000 ms)
2016-03-07 16:20:58,915 INFO [scheduler_Worker-1] [atlassian.crowd.directory.DbCachingRemoteChangeOperations] logIfNecessary updating users - (65101/104641 - 62.2%) 338487009ms elapsed
