ls -laR hangs in /

[moonchild]

on a rp8420 running 11.11

ls -laR hangs in /

Syslog has no errors.

ll / is running fine but when I am running ll -R / I face the same problem. I tried ll -R on other mount point and the result was same.

No nfs mount point exist on server.

Any ideas?
Thanks in advance

[A. Clay Stephenson]

Eventhough you say there are no NFS mount points, is it possible that the automounter is running and you are generating a mount storm?


The most likely explanation is that you have a set on symbolic links that are cross-linked and thus looping endlessly.

What happens if you do a find . -print from /? Do it get to some point and appear to hang?

[Dennis Handly]

>Clay: What happens if you do a find . -print from /? Do it get to some point and appear to hang?

Are you expecting find(1) to get in a symlink loop, or only ls -R?
$ find SUB -follow -print
SUB
SUB/SUB2
find: SUB/SUB2/FOO creates a cycle
SUB/SUB2/FOO

I was able to get ll -RL to loop but it stopped when the line was long or depth 43?:
SUB/SUB2/FOO/SUB2/FOO/.../SUB2/FOO not found

[Peter Nikitka]

Hi,

examine the point, where your command stops, very carefully. Dividing up your command like this may help:
cd /
for d in .??*
do
print === $d =====
ls -goR $d >/dev/null
# ls -goR $d
done

mfG Peter

[Ralph Grothe]

How far does -R traverse?
Do you get any listing up to a certain directory?
If so, could it be that ls cannot stat a certain filesystem because it might have become corrupt?
Do you have any stale extents in any of the mounted volumes?

[moonchild]

we just tried running it on the console and it worked fine.

We are looking into networking patches

Will keep you posted

thanks all

[Matti_Kurkela]

Is the automounter (or autofs) running with default settings?

Type "mount": if you see "/net" in the list of mounted filesystems, automounter or autofs is running. Any attempt to read the listing of /net directory causes the automounter to search for NFS shares in your local network.

If NFS access is blocked by a firewall or IPfilter that just drops the NFS packets instead of sending the proper "connection refused" messages back, any attempt to access /net/* will hang a long time.

To completely disable the automounter, first run:
/sbin/init.d/nfs.client stop

then edit /etc/rc.config.d/nfsconf and make sure both AUTOMOUNT and AUTOFS are set to 0.

MK

[moonchild]

if we use another terminal it hangs if we telnet from another hpux server and run the ll it hangs. only on console it does not.

other commands run from a remote system work fine.

root:rp8420# more exports
root:rp8420# pwd
/etc
root:rp8420# ll exp*
-rw-r----- 1 root sys 0 Nov 19 2004 exports
root:rp8420# root:rp8420# pwd



root:rp8420# ll xt*
-rw-r--r-- 1 root root 0 Aug 2 07:50 xtab
root:rp8420# cat xtab
root:rp8420#


root:rp8420# /usr/sbin/showmount -a
showmount: rp8420: RPC: Program not registered
root:rp8420#


root:rp8420# showmount -e
showmount: rp8420: RPC: Program not registered
root:rp8420# showmount -a
showmount: rp8420: RPC: Program not registered
root:rp8420# more /etc/exports
root:rp8420#


root:rp8420# cat /etc/xtab
root:rp8420# cd ..
root:rp8420# ll x*
-rw-r--r-- 1 root root 0 Aug 2 07:50 xtab
root:rp8420#

any suggestions?

[Peter Nikitka]

Hi,

it would really help, if you answered our questions.

mfG Peter

[moonchild]

when we type mount we see no /net

automount is set to 0
nfs too

once again when we run the command locally form the console it works fine.

if we use other protocols like ftp it works.

[Duncan Edmonstone]

Do you get the problem regardless of whether you telnet/rlogin/ssh into the server, or is it just with one of these products you have a problem?

HTH

Duncan

[Duncan Edmonstone]

Something else to try:

on the console:

telnet 0

To connect locally - log in

Then try command again - does it now hang?

HTH

Duncan

[moonchild]

ssh hangs as well