Invalid XML characters in indexed Word documents cause Confluence to serve up invalid XML to JIRA

I've seen two cases recently where customers have reported that searching Confluence from within JIRA gave them a "The JIRA server was contacted but has returned an error response" message instead of search results. Combing through JIRA's logs, it's reporting the following:

/rest/confluenceIssueLink/1/confluence/search [common.error.jersey.ThrowableExceptionMapper] Uncaught exception thrown by REST service
java.lang.RuntimeException: Failed to parse Confluence Remote API response
Caused by: org.xml.sax.SAXParseException: Character reference "&#7" is an invalid XML character.

I was able to narrow it down to a Word document somewhere that Confluence had indexed, and it was being returned to JIRA from the search results.

Workaround

1. Navigate to the Manage add-ons page
2. Choose to display All add-ons
3. Look for Office Connector Plugin, and then disable the Word Content Extractor module
4. Reindexing Confluence

This workaround would stop the functionality to index the Word document. Once your instance is upgraded to the Fixed Version (5.5.1), ensure that this module has been re-enabled.