Debian Bug report logs - #497572
fusesmb: Unusable with "Transport endpoint is not connected" message after some activity

Toggle useless messages

Message #5 received at submit@bugs.debian.org (full text, mbox, reply):

From: Alex <koalillo@fastmail.fm>

To: Debian Bug Tracking System <submit@bugs.debian.org>

Subject: fusesmb: Unusable with "Transport endpoint is not connected" message after some activity

Date: Tue, 02 Sep 2008 21:11:33 +0200

Package: fusesmb
Version: 0.8.7-1
Severity: important

After using fusesmb for a while, I get a "Transport endpoint is not
connected" message and fusesmb does not work until I remount the filesystem.

This seems to happen only in certain scenarios- I can use mplayer to play a
file on a network share without problems, but running eog on a share fails
immediately.

This problem seemed to begin when updating libsmbclient from 3.0.30-3

Regards,

Alex

-- System Information:
Debian Release: lenny/sid
  APT prefers testing
  APT policy: (990, 'testing'), (500, 'unstable'), (1, 'experimental')
Architecture: i386 (i686)

Kernel: Linux 2.6.26-1-686 (SMP w/1 CPU core)
Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/bash

Versions of packages fusesmb depends on:
ii  fuse-utils                    2.7.3-5    Filesystem in USErspace (utilities
ii  libc6                         2.7-9      GNU C Library: Shared libraries
ii  libfuse2                      2.7.3-5    Filesystem in USErspace library
ii  libsmbclient                  2:3.2.3-1  shared library that allows applica
ii  samba-common                  2:3.2.3-1  Samba common files used by both th

fusesmb recommends no packages.

-- no debconf information

Message #10 received at 497572@bugs.debian.org (full text, mbox, reply):

From: Dimitris Lampridis <dlampridis@gmail.com>

To: 497572@bugs.debian.org

Subject: [fusesmb] Same problem

Date: Wed, 3 Sep 2008 11:44:01 +0200

Package: fusesmb
Version: 0.8.7-1

--- Please enter the report below this line. ---

I'm also seeing the same behavior here. When I navigate to our
company's network shares using my file manager (Thunar), it is a
matter of seconds before I get a "Transport endpoint is not
connected".

Using a terminal, I can navigate to same directories and do my work, no
problems.

Kind regards,

Dimitris

--- System information. ---
Architecture: i386
Kernel:       Linux 2.6.26-1-core2duo

Debian Release: lenny/sid
  500 testing         www.debian-multimedia.org 
  500 testing         security.debian.org 
  500 testing         ftp.nl.debian.org 

--- Package information. ---
Depends              (Version) | Installed
==============================-+-==============
libc6             (>= 2.6.1-1) | 2.7-13
libfuse2              (>= 2.6) | 2.7.3-5
libsmbclient     (>= 3.0.2a-1) | 2:3.2.3-1
fuse-utils                     | 2.7.3-5
samba-common                   | 2:3.2.3-1

Message #15 received at 497572@bugs.debian.org (full text, mbox, reply):

From: Alex Corcoles <alex@corcoles.net>

To: 497572@bugs.debian.org

Subject: Workaround

Date: Sun, 21 Sep 2008 15:28:59 +0200

I'm trying now smbnetfs, which doesn't seem (so far) to have this issue, 
which might be an alternative for people experiencing this bug.

-- 
  ___    mail: alex at corcoles dot net
{~._.~}   ICQ: 66791436
 ( Y )    MSN: koalillo at fastmail dot fm
()~*~()    Y!: koalillo
(_)-(_)  http://eucaliptus.ath.cx/~alex

Message #20 received at 497572@bugs.debian.org (full text, mbox, reply):

From: Rogério Brito <rbrito@ime.usp.br>

To: Debian Bug Tracking System <497572@bugs.debian.org>

Subject: fusesmb: perfectly reproducible here

Date: Sun, 5 Oct 2008 19:32:43 -0300

Package: fusesmb
Version: 0.8.7-1
Followup-For: Bug #497572

I can perfectly reproduce this bug here with a recent kernel and things
installed from unstable (which is what I am using now).

If I issue a diff between a file that is on the smb share and on a local
drive, I *always* get this problem.

Please, let me know if further information is wanted/needed.



Regards, Rogério Brito.

-- System Information:
Debian Release: lenny/sid
  APT prefers unstable
  APT policy: (500, 'unstable')
Architecture: amd64 (x86_64)

Kernel: Linux 2.6.27-rc8-1 (SMP w/2 CPU cores; PREEMPT)
Locale: LANG=C, LC_CTYPE=pt_BR (charmap=ISO-8859-1)
Shell: /bin/sh linked to /bin/dash

Versions of packages fusesmb depends on:
ii  fuse-utils                    2.7.4-1.1  Filesystem in USErspace (utilities
ii  libc6                         2.7-14     GNU C Library: Shared libraries
ii  libfuse2                      2.7.4-1.1  Filesystem in USErspace library
ii  libsmbclient                  2:3.2.3-3  shared library that allows applica
ii  samba-common                  2:3.2.3-3  Samba common files used by both th

fusesmb recommends no packages.

fusesmb suggests no packages.

-- no debconf information

-- 
Rogério Brito : rbrito@{mackenzie,ime.usp}.br : GPG key 1024D/7C2CAEB8
http://www.ime.usp.br/~rbrito : http://meusite.mackenzie.com.br/rbrito
Projects: algorithms.berlios.de : lame.sf.net : vrms.alioth.debian.org

Message #25 received at 497572@bugs.debian.org (full text, mbox, reply):

From: Varun Hiremath <varunhiremath@gmail.com>

To: Mikhail Kshevetskiy <kl@laska.dorms.spbu.ru>

Cc: alexander barakin <alex.barakin@gmail.com>, 502129@bugs.debian.org, 497572@bugs.debian.org, control@bugs.debian.org

Subject: Re: Bug#502129: smbnetfs: Unusable with "Transport endpoint is not connected" message after some activity

Date: Sun, 19 Oct 2008 19:51:26 -0400

reassign 502129 libsmbclient 2:3.2.4-1
thanks

Hi Mikhail,

On Mon, 20 Oct, 2008 at 02:07:34AM +0400, Mikhail Kshevetskiy wrote:
> Hello Varun,
> 
> I already noticed this bug. This is libsmbclient bug appeared with samba-3.2
> release. For now (until libsmbclient will be fixed), it's better use SMBNetFS
> with libsmbclient from samba-3.0.x release.

Thanks for the info. I am reassigning this bug to libsmbclient.

Regards,
Varun

Message #30 received at 497572@bugs.debian.org (full text, mbox, reply):

From: Maciej Babinski <mbabinski@networkninja.com>

To: 497572@bugs.debian.org

Subject: patch

Date: Mon, 30 Mar 2009 15:40:00 -0500

[Message part 1 (text/plain, inline)]

Per the discussion here: <https://bugzilla.samba.org/show_bug.cgi?id=5847>,
the cause appears to be the fact that libsmbclient is not thread-safe.
While fusesmb uses locking around accesses to SMBC contexts, it uses
different locks for different contexts, and ends up making simultaenous
calls into libsmbclient -- something that is not supported. Therefore,
despite the fact that a reentrant libsmbclient would be great, I believe
the bugs lies in fusesmb. The attached patch causes fusesmb to use a single
mutex for all calls, and fixes the problem for me. Whereas without the
patch, gzip'ing a single file would cause a lockup 9/10 times, I can now
consistently gzip 10 files at once with no problems.

[fusesmb-mutex.patch (text/plain, inline)]

--- fusesmb-0.8.7/fusesmb.c	2007-09-06 19:02:44.000000000 -0500
+++ fusesmb-0.8.7.new/fusesmb.c	2009-03-30 15:27:43.000000000 -0500
@@ -53,7 +53,6 @@
 */
 
 static pthread_mutex_t ctx_mutex = PTHREAD_MUTEX_INITIALIZER;
-static pthread_mutex_t rwd_ctx_mutex = PTHREAD_MUTEX_INITIALIZER;
 static SMBCCTX *ctx, *rwd_ctx;
 pthread_t cleanup_thread;
 
@@ -129,11 +128,9 @@
 
         pthread_mutex_lock(&ctx_mutex);
         ctx->callbacks.purge_cached_fn(ctx);
-        pthread_mutex_unlock(&ctx_mutex);
 
-        pthread_mutex_lock(&rwd_ctx_mutex);
         rwd_ctx->callbacks.purge_cached_fn(rwd_ctx);
-        pthread_mutex_unlock(&rwd_ctx_mutex);
+        pthread_mutex_unlock(&ctx_mutex);
         /*
          * Look every minute in the notfound cache for items that are
          * no longer used
@@ -204,11 +201,9 @@
         {
             pthread_mutex_lock(&ctx_mutex);
             ctx->timeout = opts.global_timeout * 1000;
-            pthread_mutex_unlock(&ctx_mutex);
 
-            pthread_mutex_lock(&rwd_ctx_mutex);
             rwd_ctx->timeout = opts.global_timeout * 1000;
-            pthread_mutex_unlock(&rwd_ctx_mutex);
+            pthread_mutex_unlock(&ctx_mutex);
         }
 
 
@@ -536,17 +531,17 @@
     //    return -ENOENT;
     strcat(smb_path, stripworkgroup(path));
 
-    pthread_mutex_lock(&rwd_ctx_mutex);
+    pthread_mutex_lock(&ctx_mutex);
     file = rwd_ctx->open(rwd_ctx, smb_path, fi->flags, 0);
 
     if (file == NULL)
     {
-        pthread_mutex_unlock(&rwd_ctx_mutex);
+        pthread_mutex_unlock(&ctx_mutex);
         return -errno;
     }
 
     fi->fh = (unsigned long)file;
-    pthread_mutex_unlock(&rwd_ctx_mutex);
+    pthread_mutex_unlock(&ctx_mutex);
     return 0;
 }
 
@@ -563,7 +558,7 @@
     int tries = 0;              //For number of retries before failing
     ssize_t ssize;              //Returned by ctx->read
 
-    pthread_mutex_lock(&rwd_ctx_mutex);
+    pthread_mutex_lock(&ctx_mutex);
     /* Ugly goto but it works ;) But IMHO easiest solution for error handling here */
     goto seek;
   reopen:
@@ -575,7 +570,7 @@
             tries++;
             if (tries > 4)
             {
-                pthread_mutex_unlock(&rwd_ctx_mutex);
+                pthread_mutex_unlock(&ctx_mutex);
                 return -errno;
             }
             goto reopen;
@@ -583,7 +578,7 @@
         /* Other errors from docs cannot be recovered from so returning the error */
         else
         {
-            pthread_mutex_unlock(&rwd_ctx_mutex);
+            pthread_mutex_unlock(&ctx_mutex);
             return -errno;
         }
     }
@@ -600,7 +595,7 @@
         else
         {
             //SMB Init failed
-            pthread_mutex_unlock(&rwd_ctx_mutex);
+            pthread_mutex_unlock(&ctx_mutex);
             return -errno;
         }
     }
@@ -614,11 +609,11 @@
         /* Tried opening a directory / or smb_init failed */
         else
         {
-            pthread_mutex_unlock(&rwd_ctx_mutex);
+            pthread_mutex_unlock(&ctx_mutex);
             return -errno;
         }
     }
-    pthread_mutex_unlock(&rwd_ctx_mutex);
+    pthread_mutex_unlock(&ctx_mutex);
     return (size_t) ssize;
 }
 
@@ -632,7 +627,7 @@
     int tries = 0;              //For number of retries before failing
     ssize_t ssize;              //Returned by ctx->read
 
-    pthread_mutex_lock(&rwd_ctx_mutex);
+    pthread_mutex_lock(&ctx_mutex);
     /* Ugly goto but it works ;) But IMHO easiest solution for error handling here */
     goto seek;
   reopen:
@@ -644,13 +639,13 @@
             tries++;
             if (tries > 4)
             {
-                pthread_mutex_unlock(&rwd_ctx_mutex);
+                pthread_mutex_unlock(&ctx_mutex);
                 return -errno;
             }
             goto reopen;
         }
         /* Other errors from docs cannot be recovered from so returning the error */
-        pthread_mutex_unlock(&rwd_ctx_mutex);
+        pthread_mutex_unlock(&ctx_mutex);
         return -errno;
 
     }
@@ -667,7 +662,7 @@
         else
         {
             //SMB Init failed
-            pthread_mutex_unlock(&rwd_ctx_mutex);
+            pthread_mutex_unlock(&ctx_mutex);
             return -errno;
         }
     }
@@ -681,24 +676,24 @@
         /* Tried opening a directory / or smb_init failed */
         else
         {
-            pthread_mutex_unlock(&rwd_ctx_mutex);
+            pthread_mutex_unlock(&ctx_mutex);
             return -errno;
         }
     }
-    pthread_mutex_unlock(&rwd_ctx_mutex);
+    pthread_mutex_unlock(&ctx_mutex);
     return (size_t) ssize;
 }
 
 static int fusesmb_release(const char *path, struct fuse_file_info *fi)
 {
     (void)path;
-    pthread_mutex_lock(&rwd_ctx_mutex);
+    pthread_mutex_lock(&ctx_mutex);
 #ifdef HAVE_LIBSMBCLIENT_CLOSE_FN
     rwd_ctx->close_fn(rwd_ctx, (SMBCFILE *)fi->fh);
 #else
     rwd_ctx->close(rwd_ctx, (SMBCFILE *)fi->fh);
 #endif
-    pthread_mutex_unlock(&rwd_ctx_mutex);
+    pthread_mutex_unlock(&ctx_mutex);
     return 0;
 
 }

Message #35 received at 497572@bugs.debian.org (full text, mbox, reply):

From: Kuramshin Kamil Fidailevich <kk@tatar.ru>

To: 497572@bugs.debian.org

Subject: fusesmb: Unusable with "Transport endpoint is not connected" message after some activity

Date: Mon, 8 Jun 2009 16:41:42 +0400

Also this proplem can be resolved by using -s command line switch, that 
disables multi-threaded operation
-- 
----
С уважением,
Курамшиин Камиль Фидаилевич
Центр информационных технологий Республики Татарстан
tel: +7 (843) 2317702,
email: kk@tatar.ru
icq: 1380538; jabber id: stannum@jabber.ru

Message #40 received at 497572@bugs.debian.org (full text, mbox, reply):

From: Boris Bobrov <breton.linux@gmail.com>

To: Debian Bug Tracking System <497572@bugs.debian.org>

Subject: Re:

Date: Sun, 24 Oct 2010 16:00:50 +0000

Package: fusesmb
Version: 0.8.7-1.1
Severity: normal

The bug is still there

-- System Information:
Debian Release: squeeze/sid
  APT prefers testing
  APT policy: (500, 'testing'), (1, 'experimental')
Architecture: i386 (i686)

Kernel: Linux 2.6.32-5-686-bigmem (SMP w/12 CPU cores)
Locale: LANG=ru_RU.UTF-8, LC_CTYPE=ru_RU.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash

Versions of packages fusesmb depends on:
ii  fuse-utils                2.8.1-1.2      Filesystem in USErspace (utilities
ii  libc6                     2.11.2-6       Embedded GNU C Library: Shared lib
ii  libfuse2                  2.8.1-1.2      Filesystem in USErspace library
ii  libsmbclient              2:3.5.5~dfsg-1 shared library for communication w
ii  samba-common              2:3.5.5~dfsg-1 common files used by both the Samb
ii  samba-common-bin          2:3.5.6~dfsg-1 common files used by both the Samb

fusesmb recommends no packages.

fusesmb suggests no packages.

-- no debconf information

Message #45 received at 497572@bugs.debian.org (full text, mbox, reply):

From: Brandon <winterknight@nerdshack.com>

To: 497572@bugs.debian.org, control@bugs.debian.org

Subject: fusesmb: Unusable with "Transport endpoint is not connected" message after some activity

Date: Thu, 25 Jul 2013 01:44:54 -0700

found 497572 0.8.7-1.2
tag 497572 +patch
thanks

Adding the -s command line switch helps a lot, and makes the program
usable. However, it doesn't completely resolve the problem. Fusesmb
will still sometimes crash with Transport endpoint is not connected.

To reproduce with -s:
1) Set up ~/.smb/fusesmb.conf to use a remote smb share
2) Install storebackup with apt-get install storebackup
3) fusesmb -s mntpnt
4) cd remotefs/WORKGROUP/RemoteComputerName/ShareName
5) mkdir tmp && cd tmp
6) storeBackup -s ../LargeRemoteDirNameHere -b . -S test

LargeRemoteDirName should have lots of files and take lots of time to
copy. Something like Users on Windows should work fine.

For me, this results in Transport endpoint is not connected about 50%
of the time in about 30 seconds. If it doesn't crash within 2 minutes,
fusermount -u mntpnt and try steps 3-6 again. I've noticed that it's
more likely to crash if steps 3-6 are executed as quickly as possible.
It might help to have the destination directory (tmp in this example)
already created. It might also help to increase storeBackup options
--noCopy and/or --noCompress.

The patch provided by Maciej Babinski, however, works fine. After
patching, I've used fusesmb 7 times, and it's never crashed. I've used
storeBackup to copy files back to the remote machine as well as my
local one.

-Brandon

Message #54 received at 497572@bugs.debian.org (full text, mbox, reply):

From: gregor herrmann <gregoa@debian.org>

To: 497572@bugs.debian.org, 508452@bugs.debian.org, 728634@bugs.debian.org

Subject: fusesmb: diff for NMU version 0.8.7-1.3

Date: Tue, 22 Jul 2014 23:59:56 +0200

[Message part 1 (text/plain, inline)]

tags 497572 + pending
tags 508452 + pending
tags 728634 + pending
thanks

Dear maintainer,

I've prepared an NMU for fusesmb (versioned as 0.8.7-1.3) and
uploaded it to DELAYED/2. Please feel free to tell me if I
should delay it longer.

Regards.

-- 
 .''`.  Homepage: http://info.comodo.priv.at/ - OpenPGP key 0xBB3A68018649AA06
 : :' : Debian GNU/Linux user, admin, and developer  -  http://www.debian.org/
 `. `'  Member of VIBE!AT & SPI, fellow of the Free Software Foundation Europe
   `-   NP: Funny van Dannen: Rod Weiler

[fusesmb-0.8.7-1.3-nmu.diff (text/x-diff, attachment)]

[signature.asc (application/pgp-signature, inline)]

Message #61 received at 497572-close@bugs.debian.org (full text, mbox, reply):

From: gregor herrmann <gregoa@debian.org>

To: 497572-close@bugs.debian.org

Subject: Bug#497572: fixed in fusesmb 0.8.7-1.3

Date: Thu, 24 Jul 2014 22:18:31 +0000

Source: fusesmb
Source-Version: 0.8.7-1.3

We believe that the bug you reported is fixed in the latest version of
fusesmb, which is due to be installed in the Debian FTP archive.

A summary of the changes between this version and the previous one is
attached.

Thank you for reporting the bug, which will now be closed.  If you
have further comments please address them to 497572@bugs.debian.org,
and the maintainer will reopen the bug report if appropriate.

Debian distribution maintenance software
pp.
gregor herrmann <gregoa@debian.org> (supplier of updated fusesmb package)

(This message was generated automatically at their request; if you
believe that there is a problem with it please contact the archive
administrators by mailing ftpmaster@ftp-master.debian.org)


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512

Format: 1.8
Date: Tue, 22 Jul 2014 23:47:57 +0200
Source: fusesmb
Binary: fusesmb
Architecture: source amd64
Version: 0.8.7-1.3
Distribution: unstable
Urgency: low
Maintainer: Samuel Mimram <smimram@debian.org>
Changed-By: gregor herrmann <gregoa@debian.org>
Description:
 fusesmb    - filesystem client based on the SMB file transfer protocol
Closes: 497572 508452 728634
Changes:
 fusesmb (0.8.7-1.3) unstable; urgency=low
 .
   * Non-maintainer upload.
   * Fix "FTBFS: configure: error: Please install libsmbclient header
     files.":
     use pkg-config to find smbclient headers.
     (Closes: #728634)
   * Fix "Unusable with "Transport endpoint is not connected" message
     after some activity":
     use a single mutex for all calls to libsmbclient, patch from
     Maciej Babinski
     (Closes: #497572)
   * Fix "wrong logic to check if cache process is running":
     check that the process from the pidfile is actually running, patch from
     Stephen Gran
     (Closes: #508452)
Checksums-Sha1:
 ece22ab7a2b61d70d5798e3e2204b626a0ecff5d 1849 fusesmb_0.8.7-1.3.dsc
 2a2538bfa8ae588bb3dd0e0873b0928147e2eb15 22052 fusesmb_0.8.7-1.3.diff.gz
 c3dcb77406286a2656da59bca9f5a2c3d3c91f9b 25596 fusesmb_0.8.7-1.3_amd64.deb
Checksums-Sha256:
 db608b2303484e74a3730d757dbf622585dc26a9142c3329643b14788d303f60 1849 fusesmb_0.8.7-1.3.dsc
 4981a9384434850c79c16cf0688750fe14889a4007e96e8ec70756b8b29a3310 22052 fusesmb_0.8.7-1.3.diff.gz
 84cb8d224bf2987bf4d35bc425a9aee80839cb1ef2c4d134dc06e2a10add4548 25596 fusesmb_0.8.7-1.3_amd64.deb
Files:
 8074a99d63716c3cd97e3da8812072af 25596 utils optional fusesmb_0.8.7-1.3_amd64.deb
 bc70c939362499bcbc11f1d2703e2a0a 1849 utils optional fusesmb_0.8.7-1.3.dsc
 83a6fcca12af2fb1081c3e919bb2b0b0 22052 utils optional fusesmb_0.8.7-1.3.diff.gz

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
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=sedZ
-----END PGP SIGNATURE-----

Send a report that this bug log contains spam.