Red Hat Bugzilla – Bug 72932
no/too long timeout in autofs
Last modified: 2007-04-18 12:46:09 EDT
From Bugzilla Helper:
User-Agent: Mozilla/4.0 (compatible; MSIE 6.0; Windows 98; T312461)
Description of problem:
The first thing I tried with null was autofs. If 7.3 turning off
workstation 'sprint' with disks automounted on my test workstation
slowed down the computer to snails pace untils reboot. With null this
is better: only the application requesting data is stopped for >5
minutes. I did not wait longer to see if the shell in the terminal
where I ran 'ls /home/sprint'. On Solaris it returns with no data in
about 60 seconds.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. start NIS
2. automount a disk with ls
3. shutdown the workstation exporting the disk
4. run ls
Actual Results: shell doesn't return a prompt for >5 minuts.
Can not be interruped with ^C.
Expected Results: No files listed, shell prompt after some timeout.
The same problem arises on halt/reboot. The shutdown procedure hangs (I waited
15+ minutes) with:
nfs: server stil not responding, still trying
Sounds like a kernel nfs issue, to me.
Can you reproduce this with Fedora or one of our enterprise offerrings?
use thre intr option for your nfs mounts.