On Fri, Dec 07, 2012 at 08:29:59PM +0000, quanah@zimbra.com wrote:

Subject: Re: (ITS#7466) syncreplication causes massive explosion in database size

Closing. Issue was actually triggered by the core dump in ITS#7465, which left a read txn lock on the primary database, causing the freelist to not be able to compact. Rebuilding the database cleared the lock and things went back to normal.

Would it be possible for slapd to recover from such a case automatically on startup? If so, then I think it should. If not, then I think it is worth mentioning this in the documentation or an FAQ.

Andrew