We migrated from JIRA OnDemand to a local instance. We did a full import of the OnDemand system and then added matching LDAP accounts for each of the internal users that was imported. After synchronizing and disabling the Internal Jira account DB we are finding that multiple Issues are assigned to invalid usernames, and they appear with #1 at the end.
Mitigating factors: The OD export was a slightly newer version of Jira, but it did import into a new,empty, instance of Jira locally without errors. Also, a couple of accounts were found to have the same UID in LDAP but were corrected. Jira has been restarted and resynchronized multiple times.
For example an issue in OnDemand had:
Assigned To: George Lee
Reported By: George Lee
Now in our local instance it is:
Assigned To: George Lee
Reported By: glee#1
I can manually fix this but there are a large number of issues to correct.
Where is the mapping in the DB between the issue Assignee and Reporter? If I knew how that worked I might be able to fix it directly in the DB.
As we looked into it deeper the re-index only fixed a small portion of issue assignments. and still not clear why it was only for me in the one browser. In the end to correct all of the user assignment and reporter associations required a lot of manual effort.
For open issues we were able to use the Tools - Bulk Edit option to reassociate users.
For closed issues I had to reopen, fix, and re-close them. This wasn't critical but I didn't want to leave historical issues unsearchable by user.
Still would be nice to know where we might have been able to see the incorrect user mappings in the DB.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
You need to accept one of the answers (should be a tick to the left of them all)
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Reviewed several other tickets, all which said "DON'T muck directly in the DB for anything..." which I am heeding. Also found several references to re-indexing which as a new user I didn't know what that was.
Found it in the System - Advanced - Indexing page and ran a background re-index. Only took 36seconds on our relatively small instance. Issues are all showing correct Reporter and Assignee for me now.
Still having a problem where other users are seeing the old, incorrect users assigned. One even tried running a blocking re-index, logout/in and a different browser and the issue still exists for him.
Will update this ticket once we figure that out. Seems like another level of cache somewhere.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.