What to do when Bulk Delete does not suffice?

[RV]

Hi guys,

I need to delete about 75.000 issues from a project. I would like this delete to be permanent and irreversible (so the records are really removed from the database, not only marked as deleted).

Is this possible through the web interface? Or do I need to use SQL or a CLI in some way? I would love some guidance here, not sure what to do.

Thanks in advance,
Robbert

[Nic Brough -Adaptavist- · Accepted solution]

Do NOT try it in the database unless you know exactly what you are doing.  To help convince you of that, can you name the 18 tables you need to check and potentially remove data from?

Personally, I use REST for huge updates.  It's slow, it's clunky, but it does mean I can leave a script running gradually deleting issues from a filter.  The CLI is a good choice to help you with doing it from the outside of JIRA.

(p.s. I don't know why you've said "not only marked as deleted" - when you delete a JIRA issue, it's gone, that's it.  So a standard delete will do the job)

[Phill Fox]

As an off-piste suggestion.... How many issues are you leaving in the project when you delete 75,000 issues? If it is a relatively small number here is an alternative approach you can use which remains within the UI. Usual warnings about backups before you begin of course apply.... 1. Create a new project with the same schemes and permissions as the existing project. 2. Copy across the issues to be retained to the new project. This will retain all their history and show as a move across. 3. Check (and double/triple check) that nothing is left in the old project you want to retain. 4. Delete the old project. 5. Recreate the old project again with the same settings. 6. Move the retained issues back to the original project name. To help with this you might want to look at one of the plugins that makes copying a project structure easier. As I said at the top off-piste approach but might be quicker for you in the longrun. Phill

[RV]

Thanks Nic! I am aware of the risks when touching a DB directly, so this was not my preferred way of solving this. The 'marked as deleted' was an assumption. Some applications are designed to never actually remove data from their database, the records are just marked as deleted (which causes the DB to be ever-growing). I was not aware of the fact that JIRA actually removes things. That's a good thing. @Phill, thanks for the off-piste suggestion. This could work for me too.

[Joe Pitt]

The JIRA delete function is a real delete from the database. And the number will not be reused. Going directly into the database is NOT recommended. It can mess up other stuff. there is an API you can probably use. I never delete because there is always someone that wants it back or has questions about the deleted issue. I have 'archived' issues by moving them to another project with restricted access.