How to work with issue history in the database correctly

[Elena Oleksenko]

Hi all, 

If there is a need to change or delete an issue history (in order to remove sensitive data), how to do it correctly? 

If we talk about deleting records from changegroup and changeitem tables, should we: 

- stop Jira before doing any changes

- re-index jira after the changes were made

I saw similar questions here and in some cases it was recommended to do the described above actions. I also searched Administration Knowledge Base but did not find the clear and unambiguous instructions whether this actions required or not. Please help to understand. 

Thank you!

[Nic Brough -Adaptavist-]

Your research has been good.

You should never touch a Jira database when Jira is running, as you can corrupt it (although it's almost always a loss of the change you wanted to make, so the SQL was a waste of time, sometimes you can lose other data)

Re-indexing the affected issues is necessary.  If they're in one project, you can just re-index the project.  Or if it's not horribly urgent, and the issues are currently active, just wait until something next triggers an indexing (e.g. a new comment, edit or transition)

One other thing, it is generally safer not to delete from the database.  I usually just destroy the sensitive date, overwriting it with something like "removed for security reasons" or something innocuous. 

[Elena Oleksenko]

Thank you!