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

Re: Log level and performances

--On Friday, April 16, 2004 10:52 AM +0200 "BAILLEUX Benoît FTRD/DMI/CAE" <benoit.bailleux@francetelecom.com> wrote:

Two questions about OL performances when logging is activated.
Someone asks me recently if logging has a high effect on OpenLDAP
performances. As I have no information to answer him, I play with an OL
instance already up, and I have a strange result. The platform :
Dual processor PC, with Linux MDK 9.1, two HDD, 384 Mo RAM
OpenLDAP 2.1.5, BDB 4.2.50 (data files and log file on separate disks)
Directory loaded with two millions tiny entries, based on a very specific
schema. Excerpt of the slapd.conf file :

The results are (tests repeated 3 times) :
 - no log       : 2540 op/s (cool !)
 - loglevel=256 : 29   op/s (uh !)
 - slapd -d 256 : 2577 op/s

So I have two questions :
 - is there a problem with my configuration that makes slapd so slow when
it logs to syslog ?  - why is the throughput slightly better (and not
clearly worst) when OL logs to stdout ?


I can answer this for you --

It has to do with *where* you log the output to. This is why in my OL configuration website


I suggest that you have two disk drives, with the OpenLDAP database on one drive, and logs written to the second drive. When you are loading the directory server, it is doing a large amount of reading from disk. If you are logging, it also does a large amount of writing to disk. If you are reading and writing to the same disk, you kill your performance.

Stanford uses "loglevel 256", and we have excellent performance results from our servers, but we also have two hard drives in place, with all logs on one drive, and the DB on another drive.


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