Showing results for 
Search instead for 
Did you mean: 
Sign up Log in

Best practice for old directories after migrating to Crowd

Stephen Hodgson April 6, 2017

I recently implemented Crowd for JIRA, Confluence, and Fisheye/Crucible.  This involved moving three directories (2x LDAP and the internal directory) from each JIRA and Confluence into the same three directories in Crowd.  At this point I plan on deleting the old delegated LDAP authentication directories from JIRA and Confluence as the users coming from Crowd just naturally override the old users.

However, the internal directory is a more complicated issue as it cannot be deleted.  In JIRA, I can move the Internal directory higher than the Crowd directory and then delete the Internal directory users one by one (I plan on leaving an administrative account in case of emergencies).  This is not ideal by any means, but it works to clean up old data.

I cannot do the same in Confluence.  I moved the Internal directory higher than the Crowd directory, but I get errors deleting users "because he or she has authored content in one or more spaces".  This seems like an artificial error in this case because I was able to remove two whole directories full of users who have authored content and that content was simply mapped to the new user record from Crowd.

What have others done to clean up old data after migrating to Crowd?  I really don't want to leave old data sitting around as it only serves to complicate the solution.

1 answer

0 votes
Rising Star
Rising Star
Rising Stars are recognized for providing high-quality answers to other users. Rising Stars receive a certificate of achievement and are on the path to becoming Community Leaders.
April 10, 2017

You really need to leave an admin user in JIRA and Confluence internal directory in case the link between Application and Crowd is not working.

You could move the Crowd directory in first position.
Run a query in database to indenty all users that are in internal and in Crowd directory and do nothing with these users.
And identity all users in internal and not in Crowd, should only stay the admin user account.


Stephen Hodgson April 12, 2017

Yes, I planned on leaving at least one admin user in the JIRA and Confluence internal directories for recovery purposes.

However, doing nothing with the overlapping users doesn't resolve my original complaint that there is now a lot of duplicated data that only serves to complicate the current situation. This will only get worse over time as some accounts get disabled in Crowd but don't get changed in the no-longer-used internal directory.

It would be really nice if Atlassian had a whitepaper to cover the migration to Crowd that included some more complicated scenarios (e.g. moving multiple directories to crowd) and included what to do with the data that is no longer needed in JIRA and Confluence.

As far as resolving my issue with the extra internal users that are no longer needed, it seems like there are no options outside of fixing the data directly in the database.

Suggest an answer

Log in or Sign up to answer
AUG Leaders

Atlassian Community Events