Red Hat Bugzilla – Bug 147018
Autofs ignores proto=tcp option when it gets maps from LDAP
Last modified: 2007-11-30 17:07:06 EST
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (Macintosh; U; PPC Mac OS X Mach-O; en-US;
rv:1.7.5) Gecko/20041107 Firefox/1.0
Description of problem:
When using autofs in an LDAP environment, it seems that autofs ignores
the proto=tcp argument. We noticed this when a machine behind a
firewall tries to mount a machine outside of it. Through the
automount, with proto=tcp specified, the command fails. However, when
the mount is done by hand, with proto=tcp, the mount succeeds.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Put a machine behind a firewall that blocks udp nfs traffic
2. Set up an LDAP automount for that machine, and use the option proto=tcp
3. It will fail to automount. However, mounting by hand with
proto=tcp still works.
Actual Results: The mount is only successful if done by hand.
Expected Results: autofs should use the options specified in the map
given to it by LDAP
Here is an LDIF for an example:
# This file was generated by gq 1.0beta1 (http://biot.com/gq/)
# run by mriedel Thu Feb 3 11:24:12 2005
# subtree search on server: ldap://dapper.lanl.gov:389/
# searching below:
# version: 1
description: Space on csdteam for esm unix guys
So long as you aren't using replicated server entries, an update for this
problem will be available in RHEL 3 U5.
Chris, I believe this will be addressed by the work you are doing in get_best_mount.
I am unable to replicate this problem on my system.
Can you post the entries from /var/log/messages when autofs fails to
Could you also try the latest autofs and see if you continue to have
I am unable to replicate the problem and have not recieved any information from
reporter. Closing bug 'WORKSFORME'.