The database collation 'Latin1_General_CI_AS' and table collation 'Latin1_General_CI_AS' are not supported by JIRA.

[Yamini Sharma]

JIRA Health Analyser   is showing below error. Our SQL-Database collation is already set to latin1 then also its showing the below error.

So, my Question is

1. what is the impact over JIRA from this error.
2. What  are the steps to resolve this issue.
3. if it wont affect the Day to Day business can we live with it.

image2016-7-12 19:43:47.png

Thanks

Yamini

[Nic Brough -Adaptavist-]

It can cause problems with searching, user IDs, attachments, JIRA Software displays, Service Desk calls, and all sorts of places.

You might not run into any problems.  But you probably will sooner or later.  And it's quite likely that later versions of JIRA will simply not run when you try to upgrade (Confluence has already done this)

To fix it, you can backup JIRA, point it to a valid database, and re-import, or you can stop it, and use the database tools to change the collation.

[Chris Solgat]

Just to add to Nic's answer, (for MSSQL 2012 at least), modifying the collation to an existing database is a bit tricky as all constraint items (may include Primary Keys, Foreign Keys, and Indexes) need to be removed before the collation can be updated.  Once the collation has been updated, all of those constraints will need to be rebuilt.  Depending on the size of your JIRA instance, and whether or not you've made any custom modifications to the database itself, it's usually much simpler to stand up a new database with the correct collation set and use a JIRA Backup file to re-import the data.  When the correct collation is set at the database level, JIRA will rebuild all of the needed tables with the correct collation upon the first startup with the new database.