Hello,
I have a question concerning my openLDAP
implementation. I have recently moved to ldap for all
authentication. On the client side I am using Padl
nss_ldap 226 and the server is 2.2.17. I have
populated the directory with all the account
information adn setup equiality indices objectclass,
cn, uid, etc.
Now the issue, if I run the search (no nscd):
ldapsearch -x -h ldap01 -b "o=test.com"
"(&(objectclass=posixAccount)(uidNumber=1131)"
in a while 1 loop, I will end up eating 50% of a dual
3.0Ghz server (8G RAM).
The reason for this test, is it mimicks part of our
regression system that traverses literally thousand
different uid/gid innumerating the tests.
So my question is, how can I speed up the test (it
takes for ever compared to NIS) and how can I reduce
the hit on cpu (if I launch this twice, the machine is
completely hosed).