Red Hat Bugzilla – Bug 91416
NFS: accessing unavailable nfs partitions hangs when it shouldn't
Last modified: 2007-04-18 12:53:55 EDT
From Bugzilla Helper:
User-Agent: Mozilla/4.79 [en] (X11; U; Linux 2.4.18-19.7.x i686)
Description of problem:
I NFS mount my mail server (a Redhat 8.0 system running with the
latest patches). If I reboot the mail server, my desktop boxes lock
up. I'm mounting mail with "-o rw,bg,soft,timeo=3,retrans=3" (and
I've even tried adding "intr", although it shouldn't make any
difference). If I'm sitting in /var/mail on the client when the server
goes down, and I type "ls", my shell locks for an indefinite amount of
time (and I cannot interrupt it even though I specified "intr") before
displaying ".: Input/output error" -- as I expect it to -- (sometimes
it returns quickly, other times it never returns). Help?
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. On "client", mount "server"'s /var/mail and cd into it:
# mount -o rw,bg,soft,retrans=3,timeo=3 server:/var/mail /var/mail
# cd /var/mail
2. On "server", halt it
3. Back on "client", type ls
The amount of time it takes to respond varies greatly, and I
cannot interrupt it with ^C.
Actual Results: The system hangs for an indefinite amount of time and
(with ^C, yes I check what "intr" was with stty -a) doesn't work.
Expected Results: 2 issues:
1) I should be able to interrupt it with ^C, shouldn't I?
2) Shouldn't it return in basically the same amount of time each
time I try this?
This io errors is coming from the soft mount, try using hard,intr
as the mount options.