Managing a large data set using JCR in a production environment sometimes requires special operations with indexes stored in a file System. One such maintenance operation, called "re-indexing" consists of recreating an index by re-scanning the data. You usually need to re-index when you observe odd behaviours or many errors in the logs. There can be various reasons to re-index. For example, hardware faults, hard restarts, data corruption, migrations, or simply after an upgrade to enable new features. A re-index is usually requested on server startup or in runtime.

In eXo Platform, there are two kinds of indexes: Elasticsearch and JCR. They have different re-indexing procedures which will be the purpose of the next sections.

Match what's below in this chapter: