the index is a kind of cache for your Jira. It is not stored in your database, but on the filesystem of your Jira server. It is used for faster search of your issues, full text search and so on.
The index is a local cache of part of the database that JIRA uses for searching and reporting because database access is too slow.
I'm afraid the rest of your question does not make much sense. "if indexing is crash what are database tables to recover" - I don't know what that bit means.
