CorruptIndexException in Jira Index Manager

Jira Data Center and Server rely on a Lucene index for search, managed through the Index Manager. Documented failure conditions include IndexNotFoundException and CorruptIndexException appearing in atlassian-jira.log during startup or search operations. The Index Manager provides a Full Re-index option that rebuilds the index from the database, and a background re-indexing mode is available to reduce downtime.

There is an unresolved behavioral ambiguity around the boundary between automatic index recovery and required manual intervention. The conditions that trigger an automatic recovery attempt on startup versus a startup failure that requires explicit action in the Index Manager are not clearly defined in current documentation. Resource constraints such as CPU and I/O consumption during re-indexing and disk space requirements in Jira Home add further decision complexity for large datasets.

Under what conditions does Jira attempt automatic recovery of a corrupted Lucene index on startup? When is manual intervention via the Index Manager required after a CorruptIndexException or IndexNotFoundException? Is background re-indexing sufficient to restore search consistency, or does it require a full re-index to be initiated manually?