getcwd() returning -ENOENT???

getcwd() returning -ENOENT???

[-rb (Robert T. Brown)]

Greetings.
I have 2 clients and a server.  On one client, every 24-72 hours
I get into a situation where the shells cd'ed into
my [automounted] home directory report:

gretchen@falcon{11}pwd
pwd: cannot get current directory: No such file or directory

On the other client, it happens approximately once every 3 weeks.

(These are the same symptoms as described in 12/2000, here):
http://www.ussg.iu.edu/hypermail/linux/kernel/0012.0/0821.html
There did not appear to be any resolution, and my searches turned
up no similar problems since then.


Like the original poster, a "vdir /proc/self/." indicates:
lrwxrwxrwx    1 gretchen gretchen        0 Feb  7 15:25 cwd -> /net/home/gretchen\ (deleted)

While a "vdir" returns the full contents of /net/home/gretchen.

Also, "mount" shows the filesystem mounted via NFS:
gretchen@falcon{17}/bin/mount
/dev/hda2 on / type ext3 (rw)
     (snip)
automount(pid4490) on /net type autofs (rw,fd=5,pgrp=4490,minproto=2,maxproto=3)
ds:/u4/public/home on /net/home type nfs (rw,soft,intr,addr=192.168.3.5)

Note that /net/home is the mount point.  Originally, I suspected a problem
with the mount itself, or even the NFS server.  But, then I discovered,
interestingly enough, another shell which is cd'ed into /net/home/rbrown
actually *works* still!

The original poster indicated he rebooted to "fix" it.  Sure that
works, but I have also noticed that merely typing "cd" works too.

gretchen@falcon{23}pwd
pwd: cannot get current directory: No such file or directory
Exit 1
gretchen@falcon{24}echo $cwd
/home/gretchen
gretchen@falcon{25}cd $cwd
gretchen@falcon{26}pwd
/net/home/gretchen


(The shell is tcsh, not bash; I'm sure that's not relevant).
The client machines are running 2.4.20-8 (yea, this is the Redhat 9.0 kernel).
The server machine is running 2.4.20 (which I built).

Once every day or two, the client machine's syslog shows messages like:
Feb  4 23:16:52 falcon ntpd[4550]: synchronisation lost
Feb  5 11:04:28 falcon kernel: nfs: server ds not responding, timed out
("ds" is the nfs server and the ntp server).

The server machine's syslog shows no errors (just normal informative
messages for each NFS mount and umount request).

I've had the server and clients for years without problems.  Recent changes
include: upgraded clients to redhat 9.0, upgraded server kernel to 2.4.20,
and implemented software raid-1 on the server side's /u4 directory (the
partition which the client mounts as /net/home).

I see from the Changelog's that there are some NFS-related fixes between
2.4.20 and 2.4.24.  Not being a kernel hacker, I can't tell if any of the
fixes would address this problem.  Should I upgrade the client, the server,
or do I have to do both?

I would appreciate a CC: to my email address on any reply.
Thank you for your time.

		-rb

Re: getcwd() returning -ENOENT???

[Marcus Alanen]

-rb (Robert T. Brown) wrote:
> Greetings.
> I have 2 clients and a server.  On one client, every 24-72 hours
> I get into a situation where the shells cd'ed into
> my [automounted] home directory report:
> 
> gretchen@falcon{11}pwd
> pwd: cannot get current directory: No such file or directory

I've had this exact same thing. I start konsole and write "pwd", 
resulting in an error message. Merely cd:ing again "fixes" it, as you 
have also noted. My home directory was mounted over NFS. This was on a 
Fedore 1 machine with their kernel. The mountpoint was not automounted, 
however.

I wrote a C program whose only purpose was to print the current working 
directory using getcwd(3). Immediately after starting konsole, the 
program failed every time; after cd:ing it started working.

I can't unfortunately remember any further details like exact glibc or 
kernel version, and I've also lost any strace or tcpdump from that time 
as well. The problem was extremely persistent for a week or so, then 
suddenly stopped occurring. And I think neither the NFS server nor the 
client was rebooted in between.

You might want to try dumping the traffic and checking with ethereal 
what is actually sent across the wire, in both directions.

Marcus

getcwd() returning -ENOENT???

[Russell King]

Hi,

Can someone explain why I'm seeing the following on test12-pre7:

bash# /bin/pwd
/bin/pwd: cannot get current directory: No such file or directory
bash# vdir /proc/self/.
...
lrwxrwxrwx    1 root     root           0 Dec  7 14:52 cwd -> /net/raistlin/raistlin-v2.4/linux-ebsa285 (deleted)
...
lrwxrwxrwx    1 root     root           0 Dec  7 14:52 root -> /
...
bash# vdir
... <complete listing of directory> ...
bash# cat /proc/mounts
/dev/root / ext2 rw,noatime 0 0
/proc /proc proc rw 0 0
/dev/hda3 /var ext2 rw,nodiratime 0 0
/dev/hda5 /usr ext2 rw,noatime 0 0
/dev/hda6 /home ext2 rw 0 0
none /dev/pts devpts rw 0 0
none /dev/shm shm rw 0 0
automount(pid482) /net/flint autofs rw 0 0
automount(pid499) /net/raistlin autofs rw 0 0
raistlin:/usr/src/v2.4 /net/raistlin/raistlin-v2.4 nfs rw,v2,rsize=4096,wsize=4096,hard,udp,lock,addr=raistlin 0 0

Rebooting the machine and trying again caused this weirdness to disappear.

First time around, the network wasn't initially available when autofs
started.
   _____
  |_____| ------------------------------------------------- ---+---+-
  |   |         Russell King        rmk@arm.linux.org.uk      --- ---
  | | | | http://www.arm.linux.org.uk/personal/aboutme.html   /  /  |
  | +-+-+                                                     --- -+-
  /   |               THE developer of ARM Linux              |+| /|\
 /  | | |                                                     ---  |
    +-+-+ -------------------------------------------------  /\\\  |
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/

Re: getcwd() returning -ENOENT???

[Alexander Viro]

On Thu, 7 Dec 2000, Russell King wrote:

> Hi,
> 
> Can someone explain why I'm seeing the following on test12-pre7:
> 
> bash# /bin/pwd
> /bin/pwd: cannot get current directory: No such file or directory

Directory is unhashed. Normally it means that sucker had been deleted.

> bash# vdir /proc/self/.
> ...
> lrwxrwxrwx    1 root     root           0 Dec  7 14:52 cwd -> /net/raistlin/raistlin-v2.4/linux-ebsa285 (deleted)

Ditto.

> ...
> lrwxrwxrwx    1 root     root           0 Dec  7 14:52 root -> /
> ...
> bash# vdir
> ... <complete listing of directory> ...

Which means that it is _not_ deleted. Looks like it had been invalidated
for some reason. Try to reproduce that on -test10 - that should be before
NFS changes.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/