Upgrading Jira Data Center with faster reindexing
Jira 8.5 has an upgraded search-based subsystem engine — Lucene. Although this brings massive performance improvements, it also makes your current index incompatible with the new version. After upgrading, your nodes won’t have any index at all. You will need to reindex, there’s no running away from it. If you can’t allow prolonged downtime caused by reindexing, here’s some things you can do to speed it up.
Archiving issues for the upgrade
Archived issues aren’t included in the index, so it’s the easiest way to reduce its size. You’ll need to choose non-critical issues, as they won’t be available to users right after the upgrade. You’ll restore and reindex them later on, this time on a working Jira. We’ll show you how to use issue archiving with Jira 7.x.
Reindexing outside of the cluster
Reindex on the copy of your database while users continue to work with Jira. With a compatible and almost-ready index already in place, your nodes will only update the missing indices instead of creating the whole index from scratch.
Combined: Archiving issues and reindexing outside of the cluster
Combine both of these approaches to minimize downtime even more. Archive non-critical issues to reduce the index size, and create it outside of the cluster. Most benefits, but also a bit more complex.