We have found the problem is the overloading the master server, once we switch client to use a consumer server as primary ldap server, all problems are gone.

The Master has 16G RAM, two quad-core 3G CPU and it is binding to 400 client machines with 7000 TCP connections. I will try some performance tuning.

From:   markwu 
Sent:   Thursday, January 31, 2008 2:57 PM
To:     'fedora-directory-users@redhat.com'
Subject:        RE: Random UID not found problem

We just found that all cron jobs at 0,10,20,30,50 every hour would fail consistently and give the error in my original post, but any other minutes all cron jobs are OK.

We are using nss_ldap_226-18 and crontabs-1.10-7. I will check if UID not found problem occurs also at 10th minutes.

From:   markwu 
Sent:   Wednesday, January 30, 2008 2:42 PM
To:     'fedora-directory-users@redhat.com'
Subject:        Random UID not found problem

Some of our users get "UID xxxx not found"  message when they open a new terminal or run a rsh command, it appears a few times a day and  it is mostly just annoying message because users can continue to work as normal, however, sometimes It also causes  cron jobs to fail,

In system log, it shows,

crond(pam_unix)[9225]: could not identify user (from getpwnam(USERNAME))
crond[9225]: User not known to the underlying authentication module

We are using Fedora DS 1.0.4, and clients are RHEL 4.5 . This problem started ever since we switched into LDAP three months ago.