Dropdowns and Quick Search Not Functional After Upgrade

Brandon Viertel

We are upgrading from Jira 7.7.2 to 8.2.0 and are experiencing an issue pertaining to the following error: 

/rest/webResources/1.0/resources [c.a.plugin.servlet.DefaultServletModuleManager] Unable to create new reference LazyLoadedFilterReference{descriptor=com.atlassian.plugins.atlassian-plugins-webresource-rest:web-resources-filter (null), filterConfig=com.atlassian.plugin.servlet.filter.PluginFilterConfig@635ea5f7} io.atlassian.util.concurrent.LazyReference$InitializationException: java.lang.RuntimeException: Could not load wadl generators from wadlGeneratorDescriptions.

Jira starts, but searching and the drop-downs at the top are broken and don't display any info...

Any thoughts on this?

Thanks!

Andy Heinzer

Hi Brandon,

I have not seen this specific error before, but I would start by trying to look deeper at the $jirahome/log/atlassian-jira.log file.  Specifically I would try to look for the string 'FAILED PLUGIN REPORT' (no quotes) to start with.  If this exists, it just means that when Jira started at least one plugin did not load as expected.  It's very common to see this with upgrades since most plugins need an update to be functional in a new version, but it can also sometimes tell us about Jira system plugins that might not have loaded as expected.

Additionally you can try to look at this from the database point of view.  You can try to run the following SQL query

select * from pluginstate;

This returns a list of all plugins that Jira has flagged to be disabled.  Curious what values might exist there.

Starting with Jira 7.8 there was a new feature to improve the quick search functions, more details in Jira Core 7.8 release notes.  So it kind of makes sense to me that maybe part of this upgrade did not succeed here as expected.  It might help try to search that same Jira log file for the string 'sql exception' to see if perhaps Jira had a partial upgrade failure on a sql level and they tried to continue on.  This is largely a hunch, but this is something I would look for in the logs to try to troubleshoot this further.