What did I miss relating to MySQL and UTF-8 in the installation wizard?

[Nick Reilingh]

I just installed confluence 5.6.6 fresh on MySQL. I know our database was created with UTF-8 as the encoding (as instructed in the DB config documentation), but during the in-browser installation wizard, I accidentally skipped over a note about UTF-8. To be specific, this is right after you select your database platform (MySQL) and direct JDBC connection, there's a form for the DB connection string, and a note  that you should do something to this connection string related to UTF-8 in the event that you are using MySQL.

Can anyone tell me what this step was, and if there's something I need to do to fix it?

[Nick Reilingh · Accepted solution]

In the spirit of StackExchange-like Q&A systems (such as this), I'm going to answer my own question, now that I've basically figured out what's going on:

The note that came up in the setup wizard occurred at step 5 of the Confluence Setup Guide documentation. This displayed because we were making a direct connection to a MySQL database. To get to the point of the matter, you must dig deeper into the documentation to Configuring Character Encoding: Database Encoding, which notes specifically that the MySQL JDBC driver must have Unicode encoding set explicitly in the JDBC URL.

One more level into the docs gets us to our destination: Configuring Database Character Encoding. This contains the specific instruction for modifying the JDBC connection URL that was present in the Setup Wizard.

Finally, I will note the Troubleshooting Character Encodings page, which contains instructions for running an internationalization self-test on your database. However, in my experience this produces false-positive results, where the characters tested work even when full unicode does not. Regardless, it still contains links to configuration documentation for all aspects of character encodings for a Confluence instance, including those described here.

[Mallmann]

Hey Nick,

Every configuration needed to use Confluence with MySQL is described on our documentation about it. So I'd recommend you to check the following link out: https://confluence.atlassian.com/display/DOC/Database+Setup+For+MySQL

Let us know if you have any questions on it, alright?

I hope it helps. 

Eduardo

[Nick Reilingh]

Thanks Eduardo -- I actually added my own answer after digging into this a little deeper -- the information I was looking for was actually hidden in the character encodings documentation, and not the MySQL db setup documentation (although they are related).

[Nick Reilingh]

Small update: My hunch was that a good litmus test would be throwing a bunch of emoji characters on a page and seeing if they come back out intact. They actually _do_, but only for a short time following the page save, so I figure they must be coming from a cache until the application has to get them out of the database again (at which point they are output as {{??}}).