[389-users] RH 3-5 systems hanging requiring autofs restart to fix issue

[Charles Gilbert]

Hi everyone,

I am experiencing an issue with  my systems in that autofs, or even nscd or
crond hangs after our RH 3, 4, and 5 machines are being used for a while.
This issue is causing concern that our LDAP install is not stable obviously,
and has sent me on a goose chase to find some answers.

Errors I get in the syslog are that automount: nss_ldap: can not contact
LDAP server.

However, I am not using LDAP to store automount maps.  Our home maps are
local files that are read, and the entry in nsswitch.conf has automount:
files only.

My theory is that since we are mounting home dirs with automount, it is
trying to verify ownership and permision by contacting the server.  However,
after a time of users logging in and out on the servers, I have to log in as
root and restart autofs to allow users to continue to use the system.

I am not sure if this is an issue with tuning 389 or if it is my client
configuration.  I have read about instances where soft binds can cause
issues, and instead, people are setting their systems to hard bind and using
the nss_reconnect_  values to prevent the connection from hanging
infinitly.

Any help would be greatly appreciated!

Thanks!
Chuck
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.fedoraproject.org/pipermail/389-users/attachments/20100223/48667455/attachment.html>