Character Encoding Issue with Search in Confluence 5.1

[Charles Chappell]

Using Confluence 5.1, we're experiencing an issue when attempting to search in high character languages (Japanese in this case) where both the suggested search keyword and search box input on the search page are displaying broken characters.

メインテナンスフロー becomes メインテナンスフロー

The usual checks all come back positive for configuration being correct, and we can edit pages normally in high characters; only searching causes issues.

The server is a Windows server and is using IIS/Tomcat Connector to proxy the requests to the confluence instance.

[Joachim Ooi · Accepted solution]

Hi Charles... have you tried running Encoding Test already? It will help you to determine where the problem is coming from. Access the Encoding Test page via the <confluence base-url>/admin/encodingtest.action page on your Confluence instance. You will be required to copy and paste a line of text and submit a form. The test will take the text and pass it through Confluence, the application server and the database, and return the results.

A successful encoding test will not show you any errors.

If everything is showing fine, then you will need to check on these 3 possibilities:

1. Database side - https://confluence.atlassian.com/display/DOC/Configuring+Database+Character+Encoding
2. Application server - https://confluence.atlassian.com/display/DOC/Configuring+URL+Encoding+on+Tomcat+Application+Server
3. Web application (Confluence) - https://confluence.atlassian.com/display/DOC/Configuring+Character+Encoding

[Charles Chappell]

As it turns out... I had upgraded Confluence recently, and left out the UTF-8 URL Encoding flags in Tomcat's server.xml. (2) in the list above solved the issue for me.

Thanks for the reminder, even when we know something needs to be done, it's easy to forget sometimes in practise.