Finally ended with: Nov 2 18:00:41 ldap1 slapd[1525]: ch_calloc of 1 elems of 392 bytes failed
I've received this error when slapd was unable to allocate memory (usually by there being no free memory left on the system). This could well be caused by faulty memory chips...
Yah, could be.
this cache is related to the cachesize in DB_CONFIG - I'm assuming this uses extra memory outside the DB_CONFIG size. Is that correct?
Yes, it uses extra memory outside of DB_CONFIG.
Thought so, thanks for the clarification.
When you get a chance to experiment, what I would do is set up a server that's identical, and then run a script to load your entire database up into cache (just an ldapsearch across every entry via objectclass=* or something should do...).
That should give you an idea of your maximum memory usage.
I personally run a script on my directory servers that monitors total memory usage, and pages me when it gets too high. This was more a problem in the mid-2.2 releases when there were many memory leaks than in later releases. You may want to consider something similar. A basic concept script is on my website <http://www.stanford.edu/services/directory/openldap/configuration/> under the monitoring section.
-Dusty