We're trying to upgrade our Jira (6.0.1) to the latest version (6.2.2) and we keep getting the error "No group named {version}". After some troubleshooting, it seems that the exported xml file (entities.xml) contain free text fields (like issue descriptions) containing the character '$' that are interpreted by the importer as "group references".
This seems to be a known (?) bug that is supposed to have been fixed in an earlier version of Jira, so I think that we have missed something in the exporting process.
We're using MySQL 5.6.10 as database backend.
The problem was twofold.
1. Some user names contained '$' characters. These users had to be removed. It didn't work to just remove the '$' signs as this resulted in duplicate users (we're using a very large AD directory)
2. Other text data (summaries, descriptions, comments etc.) also contained '$' characters that had to be removed before importing was successful.
Hello Martin,
We have a documentation page regarding this problem:
https://confluence.atlassian.com/display/JIRAKB/JIRA+Upgrade+Fails+with+'NullPointerException'+Due+to+Missing+Name+Field
Could you take a look at the resolution suggested?
Benito Picarelli
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.