smbfs: mounting remote samba-filesystem leads to errors in /var/log/syslog

[Carthik Sharma]

Note: This bug was originally reported using reportbug to the ubuntu-users mailing list.

---------- Forwarded message ----------
From: Ralf Matulat <email address hidden>
To: Ubuntu Bug Tracking System <email address hidden>
Date: Sun, 23 Apr 2006 10:27:45 +0200
Subject: smbfs: mounting remote samba-filesystem leads to errors in
/var/log/syslog
Package: smbfs
Version: 3.0.22-1ubuntu1
Severity: important

-- System Information:
Debian Release: testing/unstable
 APT prefers dapper-updates
 APT policy: (500, 'dapper-updates'), (500, 'dapper-security'), (500,
'dapper-backports'), (500, 'dapper')
Architecture: i386 (i686)
Shell: /bin/sh linked to /bin/bash
Kernel: Linux 2.6.15-20-386
Locale: LANG=de_DE.UTF-8, LC_CTYPE=de_DE.UTF-8 (charmap=UTF-8)

Versions of packages smbfs depends on:
ii libc6 2.3.6-0ubuntu17 GNU C Library: Shared libraries
an
ii libcomerr2 1.38-2ubuntu1 common error description library
ii libkrb53 1.4.3-5 MIT Kerberos runtime libraries
ii libldap2 2.1.30-12ubuntu3 OpenLDAP libraries
ii netbase 4.24ubuntu2 Basic TCP/IP networking system
ii samba-common 3.0.22-1ubuntu1 Samba common files used by both
th

smbfs recommends no packages.

-- debconf-show failed

Mounting remote samba-filesystem leads to following error in
/var/log/syslog:
"Apr 22 22:14:01 buechse kernel: [4303022.782000] smb_add_request:
request [efdcfcc0, mid=16501] timed out!"

This error comes every 10-30 seconds.
The remote samba-filesystem is a Buffalo Linkstation NAS.

After mounting the remote filesystems are accessible for a while, but
then the connection gets lost and the mentioned errormessage occurs.

ls /smb/mountpoint leads to an I/O error after seconds/minutes,
df -h shows the local filesystems and than hang for a couple of
seconds/minutes before an I/O error is reported.
<Part of /etc/fstab>

//nas/milli /home/ralf/mnt/home smbfs
credentials=/home/ralf/.smbpasswd,uid=ralf,gid=ralf 0 0
//nas/movies /home/ralf/mnt/movies smbfs
credentials=/home/ralf/.smbpasswd,uid=ralf,gid=ralf 0 0
//nas/musik /home/ralf/mnt/musik smbfs
credentials=/home/ralf/.smbpasswd,uid=ralf,gid=ralf 0 0
//nas/all /home/ralf/mnt/all smbfs
credentials=/home/ralf/.smbpasswd,uid=ralf,gid=ralf 0 0

</fstab>

I have to mention that under Kubuntu Breezy it works fine for months
with the same configuration on the same NAS Device.

Now I am using cifs instead of smbfs and it seems to work for
me.

[jedie (launchpad-net-jensdiemer)]

I have timeouts, too :(

Jun 26 16:29:23 localhost kernel: [17180379.476000] smb_add_request: request [f29f2cc0, mid=21356] timed out!
Jun 26 16:29:25 localhost ntpd[4796]: kernel time sync enabled 0001
Jun 26 16:29:31 localhost kernel: [17180386.880000] smb_receive_header: short packet: 32
Jun 26 16:30:01 localhost kernel: [17180416.872000] smb_add_request: request [f29f2cc0, mid=58667] timed out!
Jun 26 16:30:01 localhost kernel: [17180416.912000] smb_add_request: request [f29f20c0, mid=58668] timed out!
Jun 26 16:30:01 localhost kernel: [17180416.944000] smb_add_request: request [f29f21c0, mid=58669] timed out!
Jun 26 16:30:01 localhost kernel: [17180416.944000] smb_lookup: find //.Trash-jens failed, error=-5
Jun 26 16:30:31 localhost kernel: [17180447.744000] smb_add_request: request [f29f20c0, mid=59477] timed out!

I used a smbfs mount in rythmbox. If i use smb://192.168.7.2/MP3, it seems to works fine.

[Shahar Weiss (sweiss3)]

I'm having a similar problem, which results in an occasional Input/output error when accessing the mounted share.
This happens at random - sometimes it works fine and sometimes I receive the error.

I have 2 samba shares being used, one of which works flawlessly at all times.

[Keith Ching (keching)]

2.6.20-15-generic #2 SMP Sun Apr 15 06:17:24 UTC 2007 x86_64 GNU/Linux

same problem, seems to occur when recursive copying directories of large files as
copying the files one at a time seems to work

input/output error
df hangs

logs have
smb_add_request: request [ffff810100ed4200, mid=51047] timed out!

my work around is to export a drive as NFS and mount it on another computer running
redhat where the smb shares can be mounted.

copy from the smbshares to the NFS on ubuntu.

[Mark Max-Hansen (k03mkm-student)]

yeah, i also have the same time-outs as jedie has in my syslog, with the prompt freezing if i do ls f.ex.

[Mark Max-Hansen (k03mkm-student)]

this means that this problem still exists in breezy. is it a gnome problem? (ie, trying to creat .Trash-$user ) ?

[Matthias (m-kaeppler)]

I can confirm this.

I am sharing files via SMB using a QNAP TS-209 Pro NAS server, and while accessing the samba shared through the local network works fine, accessing it via the internet always breaks with timeout errors:

Oct 19 10:10:43 tron kernel: [ 2311.380000] smb_add_request: request [e4873180, mid=430] timed out!
Oct 19 10:10:45 tron kernel: [ 2312.660000] smb_add_request: request [e4873c80, mid=431] timed out!
Oct 19 10:11:14 tron kernel: [ 2341.992000] smb_add_request: request [e4873080, mid=432] timed out!
Oct 19 10:11:14 tron kernel: [ 2341.992000] smb_add_request: request [e4873e80, mid=433] timed out!
Oct 19 10:12:20 tron kernel: [ 2407.496000] smb_add_request: request [e4f87ec0, mid=436] timed out!
Oct 19 10:12:50 tron kernel: [ 2437.496000] smb_add_request: request [e4f87ec0, mid=437] timed out!
Oct 19 10:13:26 tron kernel: [ 2473.512000] smb_add_request: request [e4f870c0, mid=441] timed out!
Oct 19 10:15:21 tron kernel: [ 2588.540000] smb_retry: signal failed [-3]
Oct 19 10:16:00 tron kernel: [ 2627.792000] smb_add_request: request [e4866ec0, mid=444] timed out!

My fstab entry:
//localhost/Qmultimedia /home/matthias/media smbfs user,rw,noauto,uid=matthias,credentials=/home/matthias/.smbcreds 0 0

(I'm using SSH port forwarding, in case you wonder why I mount a 'localhost' directory.)

[Mathias Gug (mathiaz)]

On Fri, Oct 19, 2007 at 08:22:08AM -0000, Matthias wrote:
> My fstab entry:
> //localhost/Qmultimedia /home/matthias/media smbfs user,rw,noauto,uid=matthias,credentials=/home/matthias/.smbcreds 0 0
>

Could you use cifs instead of smbfs ?

--
Mathias

[Matthias (m-kaeppler)]

Hi,

that doesn't seem to work at all for me. I attached the new syslogd output, for both CIFS and SMBFS (contains a couple new errors, maybe that helps). Just to be sure: I opened ports 445 and 139 for both UDP and TCP on my router *and* I tried using SSH portforwarding, so it can't be my routing table right?

Oct 23 10:40:11 tron kernel: [ 3959.544000] CIFS VFS: Error connecting to IPv4 socket. Aborting operation
Oct 23 10:40:11 tron kernel: [ 3959.544000] CIFS VFS: cifs_mount failed w/return code = -111
Oct 23 10:41:25 tron kernel: [ 4033.676000] smbfs: Unrecognized mount option noexec
Oct 23 10:41:26 tron kernel: [ 4034.992000] smb_proc_readdir_long: error=-13, breaking
Oct 23 10:41:26 tron kernel: [ 4035.036000] smb_lookup: find .Trash-matthias/.hidden failed, error=-13
Oct 23 10:44:59 tron kernel: [ 4247.764000] smb_add_request: request [e3729dc0, mid=329] timed out!
Oct 23 10:44:59 tron kernel: [ 4247.792000] smb_add_request: request [e3729ec0, mid=330] timed out!
Oct 23 10:45:00 tron kernel: [ 4248.860000] smb_add_request: request [e4920d80, mid=331] timed out!
Oct 23 10:45:00 tron kernel: [ 4248.860000] smb_lookup: find //.Trash-matthias failed, error=-5
Oct 23 10:45:29 tron kernel: [ 4277.764000] smb_add_request: request [e3729dc0, mid=332] timed out!
Oct 23 10:45:29 tron kernel: [ 4277.860000] smb_add_request: request [e3729ec0, mid=333] timed out!
Oct 23 10:45:35 tron kernel: [ 4284.156000] smb_lookup: find //.Trash-matthias failed, error=-512
Oct 23 10:45:35 tron kernel: [ 4284.160000] smb_lookup: find //.Trash-matthias failed, error=-512
Oct 23 10:45:35 tron last message repeated 2 times
Oct 23 10:45:35 tron kernel: [ 4284.164000] smb_lookup: find //.Trash-matthias failed, error=-512
Oct 23 10:45:35 tron last message repeated 2 times
Oct 23 10:45:35 tron kernel: [ 4284.168000] smb_lookup: find //.Trash-matthias failed, error=-512
Oct 23 10:45:35 tron kernel: [ 4284.272000] smbiod_handle_request: smbiod got a request ... and we don't implement oplocks!
Oct 23 10:45:36 tron kernel: [ 4284.788000] smb_get_length: Invalid NBT packet, code=15
Oct 23 10:45:48 tron kernel: [ 4296.928000] smb_add_request: request [e37298c0, mid=334] timed out!
Oct 23 10:46:05 tron kernel: [ 4314.168000] smb_add_request: request [e3729cc0, mid=359] timed out!
Oct 23 10:46:05 tron kernel: [ 4314.168000] smb_add_request: request [e3729dc0, mid=360] timed out!
Oct 23 10:46:05 tron kernel: [ 4314.168000] smb_lookup: find //.Trash-matthias failed, error=-5
Oct 23 10:46:06 tron kernel: [ 4314.748000] smb_add_request: request [e3729ec0, mid=361] timed out!
Oct 23 10:49:50 tron ntpd[5487]: synchronized to 130.149.17.8, stratum 1
Oct 23 10:51:51 tron kernel: [ 4659.800000] SMB connection re-established (-5)
Oct 23 10:55:17 tron kernel: [ 4866.312000] smb_add_request: request [e3952e80, mid=667] timed out!
Oct 23 10:55:50 tron kernel: [ 4899.344000] smb_add_request: request [e308dec0, mid=668] timed out!
Oct 23 10:55:50 tron kernel: [ 4899.388000] smb_retry: signal failed [-3]
Oct 23 10:56:20 tron kernel: [ 4929.388000] smb_add_request: request [e308d0c0, mid=669] timed out!

[Matthias (m-kaeppler)]

I made a couple of more interesting observations:

I got CIFS to work (where is the difference to SMBFS anyway, isn't that just another name for it?), but only via the SSH tunnel (listening on 445)! When mounted like that, it works. It's not fast, but it works, no more timeout errors, although I still get an occasional:

Oct 23 11:10:05 tron kernel: [ 5753.604000] CIFS VFS: Error 0xfffffff3 on cifs_get_inode_info in lookup of /.Trash-matthias/.hidden

Not sure what that means and if it's problematic at all, since as I said, I can browse my shares at least.

However, what still doesn't work at all is mounting the CIFS share directly without going through an SSH tunnel. The client just never gets a response back from the server. As I said, ports 445 and 139 are correctly forwarded and it works in the local network. It might be a server issue, but I doubt it.

[Mathias Gug (mathiaz)]

On Tue, Oct 23, 2007 at 09:20:00AM -0000, Matthias wrote:
> I got CIFS to work (where is the difference to SMBFS anyway, isn't that
> just another name for it?),

No. cifs and smbfs are different kernel modules. smbfs is no longer
maintained and should be replaced by cifs.

> However, what still doesn't work at all is mounting the CIFS share
> directly without going through an SSH tunnel. The client just never gets
> a response back from the server. As I said, ports 445 and 139 are
> correctly forwarded and it works in the local network. It might be a
> server issue, but I doubt it.

It may also be your ISP that blocks port 445 and 139, as a lot of
viruses use it to scan for windows computers.

[Steve Langasek (vorlon)]

Thank you for taking the time to report this issue and help to improve Ubuntu.

As of Ubuntu 8.04, the smbfs driver is obsoleted in favor of the cifs driver (due to a number of unresolved bugs in the smbfs driver, including this one). Therefore, I believe this issue can be considered resolved for 8.04. If you experience other issues with the cifs driver, please feel free to open new bug reports about those problems.