Where does issue keys live in JIRA

[Jig Bhakta]

Hi,

When I created a project, I realised afterwards that the issue key was misspelt (OPENUGRADE). Without realising that in our version of JIRA we could update the issue key through the UI, I went ahead and updated the database. I updated the "project" table (OPENUGRADE to OPENUPGRADE = missing P is added) but this then broke my project view and issue view, even though the issue key is shows the correctly spelt name (OPENUPGRADE-1). I then realised that this was the wrong approach so I have reverted the change in the db to update the "project" table to put back the misspelt pkey (OPENUPGRADE to OPENUGRADE = removed P) and even though I can now access my project configuration information through the Projects menu (I wasn't able to after making the initial change), the tickets still show the correctly spelt pkey (OPENUPGRADE-1). JIRA tells me that the issue does not exist but when I update the URL in the browser from OPENUPGRADE-1 to OPENUGRADE-1, I am able to access the issue.

What am I missing here?

We are using version 6.2.6 of JIRA.

Thanks in advance!!

[Nic Brough -Adaptavist- · Accepted solution]

Arg. Never touch a Jira database when it's running. Generally, 99% of people who try it even when it's not running shouldn't have because they don't understand it well enough.

Try the following, and then consider restoring from the backup you should have taken before running the SQL.

1. Stop Jira

2. Check that the database really is still correct, the shutdown has not re-written bad data into your project table.

3. Restat Jira

4. Re-index it in full (the option to lock it and reindex from scratch)

[Jig Bhakta]

Thanks Nic. Yep, it was silly of me. It sounds as if there is some caching going on here, even though the tables look the way they should?

Also, how does a full re-index differ from the background one?

Thirdly, I see that you can reindex just a project - should I give this a go instead of a full reindex? However, it doesn't state what state it will leave JIRA in, e.g. running or unavailable like a full reindex: https://confluence.atlassian.com/display/JIRA/Search+Indexing

[Nic Brough -Adaptavist-]

Yes, there's a lot of caching, that's why you need to stop and start it.

A background re-index does not delete everything and start again, unlike the full one. A project index is similar to the background one - it might fix it all, but you really should be doing a complete one.