Hi,
During performance testing with SLAMD we noticed the dncache exceeds the limit. As a result the system starts to swap a lot and the responses of slapd are very slow.
Has anyone seen that behavior before?
Here are the facts:
slapd.conf monitoring on tool-threads 4 cachesize 450000 dncachesize 450000 idlcachesize 450000 cachefree 90000
# ldapsearch -x -D "cn=xxxx" -w xx -b 'cn=database 2,cn=databases,cn=monitor' -s sub '(objectclass=*)' '*' '+' | grep -i Cache
olmBDBEntryCache: 398306 olmBDBDNCache: 482001 <<========== olmBDBIDLCache: 449999
# slapd -V @(#) $OpenLDAP: slapd 2.4.11 (Sep 11 2008 10:58:58) $ root@node3:/usr/src/redhat/BUILD/openldap-2.4.11/servers/slapd
#db_stat -V Berkeley DB 4.6.21: (September 27, 2007)
# uname -r 2.6.9-67.ELsmp
# cat /etc/redhat-release Red Hat Enterprise Linux AS release 4 (Nahant Update 6)
Cheers, Pavlos