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

Re: High volume ldap queries causing corruption in bdb backend.

--On Wednesday, November 19, 2003 11:50 AM +0000 Duncan Brannen <dbb@st-andrews.ac.uk> wrote:

set_cachesize 0 52428800 0

I'm not sure that a 50MB cachesize that is put into 0 pools is of much use. Have you tried:
0 52428800 1

I'll note that here at Stanford, we have 9 replica's that are running Solaris 8 on SunFire V120's (essentially Netra T1's). 3 of those 9 are dedicated solely to our email routing, and each receive approximately 330,000 searches a day. None of our servers have any DB corruption problems.


Quanah Gibson-Mount
Principal Software Developer
ITSS/TSS/Computing Systems
ITSS/TSS/Infrastructure Operations
Stanford University
GnuPG Public Key: http://www.stanford.edu/~quanah/pgp.html