[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: database limitations

I *think* (I'm stressing think because I'm still learning and I can't remember exactly what I read) I saw somehing about using the faster SLURPD to stream into the slower SLAPD ??? but I'm not sure and don't quote me on it.

From: Rafal Polonski <rafi@cogito.univ.gda.pl>
To: openldap-software@OpenLDAP.org
Subject: database limitations
Date: Fri, 24 Mar 2000 17:33:47 +0100


After testing OpenLDAP 1.2.8 and 1.2.9 I finished test with such results:
My database is:
Berkeley DB2 2.7.7
Database - 400 MB (does anyone uses much bigger databases?)
about 450.000 records
have very high traffic

Problem rises when I try to ldapsearch full database (looking up all records):
Often slapd just dies without errors in error log.
I tested on Linux 2.2 and FreeBSD 3.4 and situation unfortunatelly repeats.

The strange effect effect was noticed. With OpenLDAP 1.2.8 indexing of this database
was 2 times FASTER than with 1.2.9 (it took about 6 hours on Pentium III machine).
ldiff2ldbm without indexing works as fast as lighting but adding indexing for two fileds
makes it work very slowly. Is that correct?
Does anyone have samo ideas and tips concerning performance?

______________________________________________________ Get Your Private, Free Email at http://www.hotmail.com