Problem Definition

Currently the Lucene index is the same for whole JIRA index and for large JIRA instances it can be quite large. Each time reindex is running it reindexes archived projects which are not modified. Also to run search Lucene need to access whole index.

Suggested Solution

1. Mark project as archived (see: JRA-61683)
2. During next reindex project will be included in the archived shard and excluded from active shard.
3. Sharding the indexes by "active" vs. "archived"

Notes

See topic for more details: https://answers.atlassian.com/questions/10416064/what-is-the-difference-between-a-system-level-and-project-level-reindex

Workaround

None