--On Friday, September 10, 2010 8:25 AM +0200 tim stone timstone10001@googlemail.com wrote:
The index is working as designed, it's just filled with a lot of false matches which have to be explicitly tested against the filter to be weeded out. The objectclass index provided 355545 candidates in the range of 228 thru 355772. Some other search term provided (355755-112277) candidates, leaving the range from 112277 thru 355755 needing to be tested. If this search takes too long, then you need a larger entry cache.
I tried to increase the cachesize and idlcachesize to 100000 and restarted the slapd, but it did not help.
If you are using back-hdb, and using a subtree as the base (which it looked like you were doing), the first time through the search will be very slow while the cache is filled. That's a long-standing issue with back-hdb. Subsequent searches are substantially faster.
--Quanah
--
Quanah Gibson-Mount Principal Software Engineer Zimbra, Inc -------------------- Zimbra :: the leader in open source messaging and collaboration