Is there any limit in Jira and Confluence for no. of users search?

sanan sanan

Hi,

We are having a very strange situation in our Jira and Confluence setup.

Our Jira is running 7.13 and Confluence is running 6.2.3

Both our systems are integrated with central LDAP system and all our users access is authenticated with that only. 

We have 10K licenses in both the systems.

For last couple of days we are facing a situation where our new users are registered in LDAP are able to login into both Jira and Confluence because they are made member of the groups which application access. 

But for some reason they are not visible in Jira and Confluence when we search them in users. Which is making difficult for assigning issues to them.

Another strange issue is that when they have an active session i.e. when they are logged in, if at that time we try to search them, they are found. 

Can someone, please explain if he or she has seen such scenario. 

 

I repeat again, The users are able to login in jira and confluence but we can not find them in users search. but when they are logged in at that time they are visible.

 

Regards

sanan

Andy Heinzer

There is not a hard limit on the number of users in the Server products.  However I suspect that if you're using a Connected LDAP directory vs a Delegated directory type, that Jira could be having performance issues with trying to query/lookup all the possible users that might match the search terms entered in a very large environment like yours.

With Connected LDAP directories in Jira, Jira is having to sync user/group data on a regular interval (60 minutes is default).  However when you start getting above 5,000 users, you can really push the performance aspect here. And If Jira can't sync all that data in that interval of 60 minutes, sometimes the process can start over which can leave some users temporarily removed from the SQL tables that they would need to exist in to both login and be found by a lookup.  Delegated LDAP are less likely to see this problem because Jira isn't syncing the data repeatedly.  Users are only added to jira when they first attempt to sign in to Jira.  

I would be interested to see what logs are generated in $JIRAHOME/log/atlassian-jira.log when you try to recreate this problem.  It could be a problem with SQL lookup to a very large table like app_user or cwd_user, but it could also be that Jira is taking a very long time to sync all the users/groups involved.   The logs should tell us more about which performance aspect here is the area to focus on.

This is a common enough problem that we created some documentation on Reducing the number of users synchronized from LDAP to JIRA applications.