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

cache configuration constraints question



openldap community,

I would like your comments about some cache configuration. I'm sending 
attached my slapd.conf for full information about my system.

Since I have some memory constraints I needed to limit using cache 
configuration in slapd.conf file.

#Cache values
#cachesize       10000
cachesize    20000
dncachesize     3000000
#dncachesize    400000
#idlcachesize    10000
idlcachesize    30000
#cachefree       10
cachefree    10000

Since I was initially testing using only ldapseach through all DB, 
dncachesize in 3,000,000 was controlling correctly the use of memory.

Then using the same cache constraints in slapd.conf I create a jmeter 
script with a csv file that individually search for each entrance in the 
DB. This would be understand as individual ldapsearch with individual 
entrances search different from a ldapsearch for all entrances in DB.

But for my surprise the memory was consumed much faster and when around 
500K entrances were searched , much less than the ldapsearch full in 
3,000,000 cache, I already had the same memory consumption. In this way 
I had a crash in slapd since the memory was all allocated.

I would like your help to understand these memory constraints 
configuration and why memory consumption behavior was so different in 
these 2 cases.

For summary :

Case 1: Search full DB using a single ldapsearch (ldapsearch 
<parameters> filter=* )

Case2: Multiple searches to all entrances in DB using like multiple 
ldapsearch (ldapsearch <parameters> filter=<unique DB>)

Is there any difference between these two cases based in cache limitation?

Maybe I've some incorrect understanding about cache configuration.

Best Regards,

Rodrigo.