Hi Atlassian,

When i search issues in JIRA i often note CPU spikes on the server, especially on a VMWare guest. I guess this is caused by the Lucence search engine.
For this reason I wondered if indexing and searching could not be offloaded to a remote dedicated index server like Solr. Probably it would not be too difficult to use Solr instead of Lucene since Solr is a superset of Lucene regarding the search query syntax.

Another advantage would be that one can better scale with distributed Solr servers and also one obstacle towards a clustered JIRA installation would also be taken since the the index files are not local to a JIRA server anymore.

Probably one could also address the problem that JIRA isn't available during indexing but i must confess that i currently don't know how to do that with Solr. For large installations it is really annoying to see JIRA unavailable while it is reindexing since a lot of users are affected.

What about this idea?

Best regards,
Dieter