On 06/22/2014 07:28 PM, John Jason Jordan wrote:
> It's hard for me to remember anything in all my years of using
> computers that is as big a PITA as NFS.

Is this NFS with hard mounts or autofs? If you continue to have problems 
and you are using /etc/fstab and NFS with hard or soft mounts, you may 
want to look at autofs.

<clipped>

>>From the command line on the client the mount command executes without
> error and df -h says the share is mounted. However, when I try to
> umount the share I get error messages that the share is busy. I even
> tried umount -f, but it still says the share is busy and won't umount
> it. And the terminal window hangs if I try to cd to the share; that is,
> I can cd right up to the folder I want to see, but just typing the
> first letter of the folder hangs the terminal window, as does the ls
> command. That probably means something, but Google won't tell me what.

Basically, it sounds like the kernel became confused and thinks it knows 
how to talk to the device, but it really doesn't know any longer. This 
could a network hiccup or timeout, or a routing table error, or just 
about anything else you can make up.

The hang you see is because the kernel hasn't been given a timeout (or a 
reasonably short timeout) and is normal behavior.

Look at man for mount and look at the '-l' (little ell) flag. I dealt 
with this at one point and *think* this was what I used for clearing an 
automount issue:

# umount -l /mount/point
# cd /mount/point

This performed an unmount of the share essentially ignoring errors, and 
then changing to that directory remounted it to be available for my users.

<clipped>


> Suggestions for things to poke at would be welcome.

Hopefully this will keep you up and running without a reboot on either 
the client and/or the server.

dafr



_______________________________________________
PLUG mailing list
[email protected]
http://lists.pdxlinux.org/mailman/listinfo/plug

Reply via email to