Red Hat Bugzilla – Bug 152258
Autofs looks for spurious /net mounts
Last modified: 2008-08-02 19:40:32 EDT
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.6) Gecko/20050323 Firefox/1.0.2 Fedora/1.0.2-1.3.1
Description of problem:
NFS mounts in /net are hanging for a while, owing to some strange lookups
by the auto.net script:
Mar 26 20:42:58 charlus automount: >> /usr/sbin/showmount: can't get address for 192.168.0.10/maxtor250/libpthread.so.0
Mar 26 20:42:58 charlus automount: lookup(program): lookup for 192.168.0.10/maxtor250/libpthread.so.0 failed
Mar 26 20:42:58 charlus automount: failed to mount /net/192.168.0.10/maxtor250/libpthread.so.0
Mar 26 20:42:58 charlus automount: >> /usr/sbin/showmount: can't get address for 192.168.0.10/maxtor250/i686
Mar 26 20:42:58 charlus automount: lookup(program): lookup for 192.168.0.10/maxtor250/i686 failed
Mar 26 20:42:58 charlus automount: failed to mount /net/192.168.0.10/maxtor250/i686
Mar 26 20:43:20 charlus automount: >> /usr/sbin/showmount: can't get address for 192.168.0.10/maxtor250/libattr.so.1
Mar 26 20:43:20 charlus automount: lookup(program): lookup for 192.168.0.10/maxtor250/libattr.so.1 failed
Mar 26 20:43:20 charlus automount: failed to mount /net/192.168.0.10/maxtor250/libattr.so.1
This means than running 'ls' in the /net/192.168.0.10 directory hangs for a couple of minutes. Changing to one of the directories inside the mount works,
most of the time, without delay.
This has only started happening in the past day or so, which means it must have come in one of the recent updates.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. change to a /net mounted directory
Actual Results: That login hangs for several minutes
Expected Results: The contents of the remote directory should be listed.
I suspect this has been addressed in a later version of the package. I'll work
to get a newer version into FC3. If you are so inclined, you could grab the
srpm from rawhide and build that. Otherwise, I'll let you know when a new
version is available.
You can find a test version of the fc3 autofs package at the URL above. Please
try it out and let us know if it fixes your problem.
Have just tried that new version and the problem is still there, e.g.:
7539 pts/5 S 0:00 /bin/sh /etc/auto.net 192.168.0.10/libpthread.so.0
7540 pts/5 S 0:00 /usr/sbin/showmount --no-headers -e
7541 pts/5 S 0:00 sort +0
7542 pts/5 S 0:00 awk -v key=192.168.0.10/libpthread.so.0 -v opts=-fstype=nf
Can you post your /etc/auto.master, the output from '/etc/auto.net 192.168.0.10'
& your /etc/auto.net file.
The command output:
/home/adam 192.168.0.10:/home/adam \
Created attachment 113350 [details]
This sounds like the buggy negative dentry caching problem. This has been fixed
in recent kernels. Can you re-test with the latest bits? I'll attach the patch
that fixed things.
Can't find the patch kicking around on my system anymore. Here is a link to a
mailing list post with the patch:
Closing due to no response. Should be fixed in recent kernels.
There's still a problem with the latest FC5 updates - I have to restart the
autofs service in order for it to work.
OK, can you collect some debugging information? Take a look at the section
entitled "Filing bug reports" at the following URL:
The debug logs are really what I'm looking for.
Actually, I take that back - it does seem to be working now.
Thanks for taking a look.
I'm actually having other problems with autofs now, listed in bug #271665.