[Date Prev][Date Next]
Re: Outstanding issues between 2.0.7 and Linux?
We upgraded from 1.2.11 to 2.0.7 (both threaded) on RH6.1 and changed our
backend from GDBM to BerkeleyDB3.1 at the same time. A benchmark query
(ldapsearch memberofgroup=cn=All Staff,ou=Groups,o=MyOrg, c=US) shows
execution time jumps dramatically:
1.2.11(gdbm): 1:20.06 elapsed
2.0.7 (berkeley3.1): 7:55.60 elapsed
Same server, same client, same indexes. I haven't tested 2.0.7 with a gdbm
What could be the explanation?
----- Original Message -----
From: "Carl Litt" <firstname.lastname@example.org>
Sent: Wednesday, February 07, 2001 11:41 PM
Subject: Outstanding issues between 2.0.7 and Linux?
> I know the answers to these for 1.2.x, but I'm deploying a new 2.0.7
> installation and I wonder if the answers have changed (on the
> Linux platform).
> Is threaded 2.0.7 slapd still a bad idea under Linux? I remember
> issues with 1.2.x and Linuxthreads, but I thought I heard that 2.x
> would overcome this. (I've got a nice dual PIII/800 and I'd hate to
> see half of it go to waste).
> Is Berkeley DB 3.1 much better than GDBM when considering performance
> and reliability? I notice it's recommended, but why?
> Carl Litt
> Network Administrator
> Execulink Internet