samba - May 2000 - Errors with smbfs/smbmount

On Wed, 24 May, Todd A. Green wrote:

	I've been working with smbfs/smbmount the last few weeks to allow my
	linux log processing server access to the logs on our NT servers.
The
	problem is that after a period of time all the smbfs mounts become
	invalid with either one of two errors.

Yes. I have exactly the same thing. 'Mount' shows them still mounted, 
but 'ls -l' on the mount point shows nothing there. This is driving me
crazy.

	When you do a 'df -k' to view
	the mounts you see the following:

	samba:~# df -k
	Filesystem           1k-blocks      Used Available Use% Mounted on
	/dev/hda1              2071384    779284   1186876  40% /
	/dev/hda6             10752392     74876  10131308   1% /usr/local
	df: /servers/server1.mydomain.net: Broken pipe
	df: /servers/server2.mydomain.net: Broken pipe
	df: /servers/server2.mydomain.net: Broken pipe
	df: /servers/server2.mydomain.net: Broken pipe

	samba:~# df -k
	Filesystem           1k-blocks      Used Available Use% Mounted on
	/dev/hda1              2071384    779284   1186876  40% /
	/dev/hda6             10752392     74876  10131308   1% /usr/local
	df: /servers/server1.mydomain.net: I/O Error
	df: /servers/server2.mydomain.net: I/O Error
	df: /servers/server2.mydomain.net: I/O Error
	df: /servers/server2.mydomain.net: I/O Error

	I have a script which checks these mounts once every thirty minutes
and
	if they have errors they are unmounted and re-mounted.

I have the same kludge! Fixes the symptom but not the cause.

	The script emails me everytime it has to remount so I have some
times as
	to when it is happening.

	2000/05/23:  01:30, 02:00, 04:00, 06:00, 08:00, 10:00, 12:00, 14:00,
	16:00, 18:00, 19:00, 20:00, 22:00
	2000/05/24:  00:00, 02:00, 02:30, 04:30, 06:30, 08:30

	Since the script only runs every half-hour I don't know exactly when
the
	mounts are going sour.

	I'm running samba-2.0.7,  on a Redhat 6.1 box with a 2.2.14 kernel.
	Here's the configure statement I used to build it ./configure
	--with-smbmount --with-syslog

	Note: there are NO entries in any of the syslog files with regard to
	samba.

	I dug into the archive and the last messages I found pertaining to

Are these Samba mailing list archives? If so please, someone send me a URL.

	smbmount were from November 1998.  They talked about the NT servers
	timing out 'idle' smb mounts.  I'm wondering if this is the case
and
if
	so is there anything I can do about it.

	Only other thing of interest is that when the drives are mounted
	/sbin/mount.smbfs is left running in the background.  One for each
mount
	point.  Is this normal behavior?

It is normal on my system. For each remote mount, 'ps -ax' typically
shows:

smbmount //<server>/<share> <password> -U <username>  -W
<workgroup> \
-c mount ?/mnt/<mountpoint> -g <group> -f 775 -d 7

	If this is the wrong forum for this question please let me know
where to
	post the questions.

I'm new to this list (this problem is what prompted me to subscribe). I have

not received the normal trickle of messages that I would expect for a
mailing 
list. Instead, I received the message above in a digest of 15 or so. Is this

normal samba list behaviour, or have I done something stupid when 
subscribing? I would really appreciate an answer on this - even if nobody
knows the answer to the "disappearing NT shares" problem.

Thanks & Regards
Martin Munt

Draugen Project, SubSea Offshore Limited
Rockwater Building, Stoneywood Park, Dyce, Aberdeen, AB21 7DZ
Tel: +44 1224 795433  Fax: +44 1224 795582
mailto:Martin.Munt@Halliburton.com

>Are these Samba mailing list archives? If so please, someone send me a URL.
http://lists.samba.org/
>I'm new to this list (this problem is what prompted me to subscribe). I
have
>not received the normal trickle of messages that I would expect for a
>mailing 
>list. Instead, I received the message above in a digest of 15 or so. Is
this
>normal samba list behaviour, or have I done something stupid when 
>subscribing? I would really appreciate an answer on this - even if nobody
>knows the answer to the "disappearing NT shares" problem.
Actually this is my first post to the Samba list.  Like you I subscribed in
the hope of finding a solution to this problem.  As to the digest messages I
just got my first one today so I suppose that's how it works.

Anyway if I find out anything I'll let you know.

> 	I'm running samba-2.0.7,  on a Redhat 6.1 box with a 2.2.14 kernel.
[snip]
> 	Only other thing of interest is that when the drives are mounted
> 	/sbin/mount.smbfs is left running in the background.  One for each
> mount
> 	point.  Is this normal behavior?
> 
> It is normal on my system. For each remote mount, 'ps -ax'
typically shows:
Yes, the smbmount/mount.smbfs process hangs around. If smbfs "detects"
that the connection is down it calls on smbmount to open a new connection
to the server.
> 
> smbmount //<server>/<share> <password> -U
<username>  -W <workgroup> \
> -c mount ?/mnt/<mountpoint> -g <group> -f 775 -d 7
This syntax looks older than samba 2.0.6. Do you get the same problem if
you try a 2.0.7? There is a known problem with smbmount dying when it
shouldn't in some versions, that bug was fixed in 2.0.6 (or maybe it
wasn't until 2.0.7, but I do know that it is fixed now :).

I don't know that this will fix your problem, obviously a similar problem
is still around in 2.0.7. Personally I have not seen this since problem
for a while (ie 2.0.5, 2.0.6 or whatever it was).

> I'm new to this list (this problem is what prompted me to subscribe). I
have
> 
> not received the normal trickle of messages that I would expect for a
> mailing 
> list. Instead, I received the message above in a digest of 15 or so. Is
this
The list is also available in non-digest form. See the end of:
http://lists.samba.org/

But for some reason the default is digest form.


Offtopic:
To me the "indent-previous-response" is a bit confusing. Especially
when
the lines wrap (as it does for 'mount' above). Anything wrong with plain
old '> '?

/Urban