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

scaling openldap for performance

Hey there Fellas,

  Having tread through the archives for a long time i
have concluded with some load/performance tests.

 Openldap scales *wonderfully* with respect to read
performance. Towards testing the same for write
performance i performed the following:

> With Zero think time we hit openldap with 20
simultaenous users. The db has 1million records.  5
mins thro the test i have a seg fault. I try
restarting the same but the server refuses to re-read
the db unless i re-create the same. Find this very
strange. Can i avoid this In-consistency????

> Me brought this DB size to 100 Addressbook records
and pounded the server with 20users and Zero think
time. Still server crashes.

> Brought the users to 10 users and 100 records.
Bombarding the server with Zero think time. It now
survives and pulls through the test.

 HW Arch is : HP BLade20p, 2 CPU, 1GB RAM, RHEL AS2.1,
RAID 0+1 on a SAN.

 Am quite confused to be true.  Logging is disabled
cause we found its over head is simply too much and
hinders response times. 

 Lemme know of any runtime/compile time options
towards boosting performance please. Thanks in

( >-                                           -< )
/~\    ______________________________________   /~\
|  \) /    Scaling FLOSS in the Enterprise   \ (/ |
|_|_  \        trevorwarren@yahoo.com        / _|_|

Do you Yahoo!?
Friends.  Fun.  Try the all-new Yahoo! Messenger.