Why won't my database collation pass the health check?

[JConstantine]

I've set my database collation to utf8mb4_bin which is supported by the latest version of JIRA:but I still can't pass the health check:
What might be the problem here?

[Nic Brough -Adaptavist-]

The database does not appear to be set to utf8mb4_bin - in the result line, it is reporting what the database has told it the collation on the Jira database is.

You'll need to change the collation of the data, as well as the database setting.

[JConstantine]

@Nic Brough -Adaptavist- I did everything from here:
https://confluence.atlassian.com/kb/how-to-fix-the-collation-and-character-set-of-a-mysql-database-744326173.html

and I did this as well:

To fix this, the database collation needs to be corrected. The best and recommended way to achieve this is:

1. Create a new database with the required collation as per the appropriate documentation in Connecting JIRA to a Database.
2. Follow our Switching Databases to migrate JIRA from the old databases (with the incorrect collation) to the new one, with the correct collation.

but I still can't pass the health check.
Losing data for me is not an option.
Here's the screenshot from my MySQL server:

[Nic Brough -Adaptavist-]

The problem with this is that whatever database you are connecting to is telling Jira that it is set to utf8, not utf8mb4.  Why is it reporting it like that?

[JConstantine]

@Nic Brough -Adaptavist- I've no idea. I didn't have this problem when I was using Jira 8.6.x and MySQL 5.6.47. I did everything from the documentation but it didn't solve my problem :(

[Nic Brough -Adaptavist-]

The database itself (not the service, the individual database within it) is reporting that you created the database with utf8_bin, so I suspect you used the old collation again and missed creating the new database with utf8mb4_bin

[JConstantine]

@Nic Brough -Adaptavist- I created my new database like this:
CREATE DATABASE <dbname> CHARACTER SET utf8mb4 COLLATE utf8mb4_bin

and then I used mysqldump to transfer the old database to the new one.

Here's the screenshot:

As you can see everything is set correctly.

[Nic Brough -Adaptavist-]

I assumed you did, but we need to find out why your database is telling Jira that it is utf8 rather than utf8mb4.  Are you sure Jira is looking at the right database?

[JConstantine]

Yes. I'm sure. Here are the screenshots of Jira Configuration Tool and dbconfig.xml file:

[Nic Brough -Adaptavist-]

There's one last long-shot - could you

• Tell us what version of Jira it is (you mentioned 8.6 earlier, but I suspect you're now higher than that?)
• Tell us what the database version is (not via Jira, but directly on the database command line (don't trust everything Jira is telling you!)
• Confirm you're running everything on one single machine - Jira, the database service, the mysql command line that you've used to show us stuff and the jira config tool
• Check the database driver file - what version did you download and add to the jira directory?  (Again, don't trust Jira to tell you)

[JConstantine]

@Nic Brough -Adaptavist- 

• Jira Core 8.12.2, Jira Software 8.12.2, Jira Service Desk 4.12.2
• 
• Everything is running on a single machine
• I downloaded mysql-connector-java-8.0.21 driver and put it into <JIRA installation directory>\lib\

[Nic Brough -Adaptavist-]

@JConstantine  - I wish other people would answer my debugging questions as clearly and consistently as you have.

Annoyingly, despite the info being absolutely what I asked for, it does not help.  You have done everything absolutely right, and you've shown us what you've done which rules out everything I can think of that might cause this warning to be thrown.

All I have left to offer is:

• I want this to go to Atlassian as a support call / bug report.  The health-check is somehow inaccurate, so either there's a bug in their code or something really obscure that I am missing.  Either way, it has to be support from them as we don't know exactly what the code is doing to get to this (inaccurate) error report
• Do not worry about it generally.  Collation problems affect search, sort, and "advanced" storage of characters, but while, yes, you may lose data, most humans will be able to cope with it.  You're not looking at "can't find anything", but "cannot search for human names that contain Turkish characters when collation is set to Swedish" type things.  The computers might fail us on this, but the data is intact and most humans reading it will fill in the gaps.

[JConstantine]

@Nic Brough -Adaptavist-Got it. Do I have to create that bug report or will the Atlassian support team do that for me?
And one more thing. Even though those health checks have been successfully passed, JIRA keeps showing these warnings:

[Nic Brough -Adaptavist-]

You'll need to raise it as a support request (I'm not an Atlassian, and there's no point me raising it when I won't be able to get into your systems to find answers to questions they may ask)

The application link failure means you've got one or more application links configured to systems that are no longer responding when Jira pings them.  It won't do any harm, but you may want to delete the broken links if you don't need them any more, or recreate them if you do.

Gadget feed problem usually means your base url is set incorrectly, or that the current server cannot reach itself (the base url) over your network.  Again, won't do any harm to your data, but some things won't work properly until it is corrected.

[JConstantine]

@Nic Brough -Adaptavist-The thing is that I have recreated the links and the Base URL and as a result of that JIRA has passed the health check on those problems, but those notifications are still showing up. Anyways, I got your point.
When I'm trying to request support:
II get this message showing up:
Is there another way to raise the request?