Roland Koebler | 20 Feb 13:34 2015

sshfs, hardlinks and use_ino


I think I've found a bug in SSHFS:

- I can create hardlinks over SSHFS, but by default, stat then reports
  different inodes for the original file and the hardlink!
  (This behavior is very strange by itself, and I would consider it a
  bug by itself, since it breaks the definition of a hardlink.)

- "-o use_ino" should fix this, but when I try to mount with this
  option, the mount seems to fail silently (no error-message, mounted
  according to "mount", but there are no files).

This should be fixed, so that either hardlinks work correctly (and stat
reports the same inode-numbers for hardlinks), or that sshfs fails when
trying to create a hardlink.

Note: This e.g. means that you cannot use mutt with maildirs over sshfs
      anymore, since mutt tries to hardlink a file and then compares the
      inodes; and if they don't match, mutt will hang/loop and create
      thousands (or more) of hardlinks until mutt is killed.
      (Related Debian-bug-number: #778801)

Tested with SSHFS on Debian (both Wheezy/stable and testing):
- SSHFS version 2.4
  FUSE library version: 2.9.0
  fusermount version: 2.9.0
  using FUSE kernel interface version 7.18
- SSHFS version 2.5
  FUSE library version: 2.9.3
(Continue reading)


gnome-keyring daemon fails on SSHFS

Hi, is it possible for this bug to get solved from the SSHFS side instead
of the gnome-keyring side?

I don't know if the GNOME developers are willing to fix it in gnome-keyring
if they consider it a bug in SSHFS, but if it is a limitation in SFTP and
can't be fixed in SSHFS, we'll have a better chance of convincing them...

Thank you!
New Year. New Location. New Benefits. New Data Center in Ashburn, VA.
GigeNET is offering a free month of service with a new server in Ashburn.
Choose from 2 high performing configs, both with 100TB of bandwidth.
Higher redundancy.Lower latency.Increased capacity.Completely compliant.
leecholho | 26 Nov 01:53 2014

help: sshfs error

i use sshfs and i get error message . this is bug?
And this is connection command.sshfs  root <at> staging:/opt/XXXX/current/app/webroot
/opt/XXXX/current/app/webroot -o IdentityFile=/root/.ssh/sshfs,allow_other,uid=48,gid=48
-odebug,sshfs_debug,loglevel=debug -f
this is error logLOOKUP /img/city/T00018/T00018_xl.jpggetattr /img/city/T00018/T00018_xl.jpg  
NODEID: 18914   unique: 49919025, success, outsize: 144unique: 49919026, opcode: READDIR (28), nodeid:
18900, insize: 80, pid: 18043   unique: 49919026, success, outsize: 16unique: 49919027, opcode:
RELEASEDIR (29), nodeid: 18900, insize: 64, pid: 0   unique: 49919027, success, outsize: 16unique:
49919028, opcode: GETATTR (3), nodeid: 18904, insize: 56, pid: 18043getattr /img/city/T00010   unique:
49919028, success, outsize: 120unique: 49919029, opcode: OPENDIR (27), nodeid: 18903, insize: 48,
pid: 18043   unique: 49919029, success, outsize: 32unique: 49919030, opcode: READDIR (28), nodeid:
18903, insize: 80, pid: 18043getdir[0][12799440] OPENDIR  [12799437]         STATUS       32bytes (0ms)  [1
 2799438]         STATUS       32bytes (0ms)  [12799439]         STATUS       28bytes (0ms)  [12799440]         HANDLE       17bytes
(0ms)[12799441] READDIR[12799442] READDIR  [12799441]           NAME      815bytes (0ms)  [12799442]         STATUS       32bytes
(0ms)[12799443] READDIR[12799444] READDIR  [12799443]         STATUS       32bytes (0ms)  [12799444]         STATUS      
32bytes (0ms)*** glibc detected *** sshfs: double free or corruption (out): 0x00007f34440be390
***======= Backtrace: =========/lib64/[0x7f344c37f166]/lib64/[0x7f344c381c93]sshfs[0x4046a4]sshfs[0x406872]sshfs[0x40b384]/usr/lib/[0x7f344cfd5065]/usr/lib/[0x7f344cfdc009]/usr/lib/[0x7f344cfe0e16]/usr/l

Memory map: ========00400000-00410000 r-xp 00000000 08:01 340943                            
/usr/local/bin/sshfs0060f000-00610000 rw-p 0000f000 08:01 340943                            
/usr/local/bin/sshfs00610000-00611000 rw-p 00000000 00:00 001765000-0180a000 rw-p 00000000 00:00
0                                  [heap]7f3428000000-7f3428105000 rw-p 00000000 00:00 07f3428105000-7f342c000000 ---p 00000000
00:00 07f342c000000-7f342c1bd000 rw-p 00000000 00:00 07f342c1bd000-7f3430000000 ---p 00000000
00:00 07f34335ff000-7f3433600000 ---p 00000000 00:00 07f3433600000-7f3434000000 rw-p 00000000
00:00 07f343
 4000000-7f34341f0000 rw-p 00000000 00:00 07f34341f0000-7f3438000000 ---p 00000000 00:00
07f3438000000-7f3438190000 rw-p 00000000 00:00 07f3438190000-7f343c000000 ---p 00000000 00:00
07f343c000000-7f343c021000 rw-p 00000000 00:00 07f343c021000-7f3440000000 ---p 00000000 00:00
(Continue reading)

Transport endpoint is not connected

i use sshfs to connect another server in local network.

this is sshfs process list.
[lee <at> adm ~]$ ps aux | grep sshfs
root 2868 0.0 0.0 60344 3812 ? S Sep30 28:20 ssh -x -a
-oClearAllForwardings=yes -oIdentityFile=/root/.ssh/sshfs -2
root <at> staging -s sftp
root 2869 0.1 0.0 625280 4188 ? Ssl Sep30 96:58 sshfs
root <at> staging:/opt/ /opt/
-o IdentityFile=/root/.ssh/sshfs -o allow_other -o uid=48 -o gid=48
root 10108 1.6 0.0 60344 3884 ? S Nov18 23:46 ssh -x -a
-oClearAllForwardings=yes -oIdentityFile=/root/.ssh/sshfs -2
root <at> staging -s sftp
root 10109 2.2 0.1 702568 15604 ? Ssl Nov18 30:55 sshfs
root <at> staging:/opt/
/opt/ -o IdentityFile=/root/.ssh/sshfs -o
allow_other -o uid=48 -o gid=48

webroot mount is always broken(Transport endpoint is not connected) But
View mount is not broken naver(at least nothing)
i used option reconnect and ServerAliveInterval, BUT result is same.
i traced sshfs pid(10108) and log is this.

[lee <at> adm ~]$ sudo strace -p 4887
elect(7, [3 4], [5], NULL, {10, 0}) = 1 (out [5], left {9, 999998})
write(5, "\0\0\0\34e\1l0\2\0\0\0\1\0\0\0\vEnd of file\0\0\0\0"..., 64) = 64
select(7, [3 4], [], NULL, {10, 0}) = 1 (in [4], left {9, 966952})
read(4, "", 16384) = 0
(Continue reading)

bzk0711 | 21 Oct 11:44 2014

[Bug] Mount freezes on specific mount point

Hi there,

since two days, I have been experiencing a very strange bug with sshfs. It seems to be related to an
openssl/openssh update (since it worked before and that's the only thing which changed). However with
downgrading I was not able to isolate the issue, neither on Gentoo nor Arch, and had the issue on both systems.

I don't know if that's related, but I have in my .ssh/config:
Host sks
  HostName saunaserver
  ServerAliveInterval 10

saunaserver is a hosts entry which maps to a LAN IP. 

What I observe:
bzk <at> silencio ~ $ mkdir mnt/
bzk <at> silencio ~ $ sshfs sks:/mnt/raid mnt/
bzk <at> silencio ~ $ ls mnt/
data  literature  lost+found  music  rcs  roms  share  software  torrents  video
bzk <at> silencio ~ $ fusermount -u mnt/
bzk <at> silencio ~ $ mkdir sks
bzk <at> silencio ~ $ sshfs sks:/mnt/raid sks/
bzk <at> silencio ~ $ ls sks
data  literature  lost+found  music  rcs  roms  share  software  torrents  video
bzk <at> silencio ~ $ fusermount -u sks
bzk <at> silencio ~ $ mkdir data/media/music/server
bzk <at> silencio ~ $ sshfs sks:/mnt/raid data/media/music/server/
bzk <at> silencio ~ $ ls data/media/music/server/
data  literature  lost+found  music  rcs  roms  share  software  torrents  video
bzk <at> silencio ~ $ fusermount -u data/media/music/server/
bzk <at> silencio ~ $ mkdir data/media/music/sks
(Continue reading)

Jasper Lyons | 14 Oct 18:06 2014


Hey guys

I’m running a macbook pro

OSX: 10.9.5
2.6Hz i5
8GB ram (1600Mhz DDR3)

SSHFS version 2.5 (OSXFUSE SSHFS 2.5.0)
OSXFUSE library version: FUSE 2.7.3 / OSXFUSE 2.7.1

When I try and run 

	sshfs user <at> host ~/amount/point -o allow_other,reconnect,volname=AName 

from within a launchctl launch agent I get no error but no mounted drive either.

If I replace the command with

	sshfs user <at> host ~/amount/point -o debug,allow_other,reconnect,volname=AName

again from within a launchctl launch agent I get all of the debugging output to stdout and the remote fs is
mounted at ‘~/amount/point’

This seems like pretty sketchy behaviour, anyone have any ideas why this might be the case?

Comprehensive Server Monitoring with Site24x7.
(Continue reading)

Reuben Thomas | 3 Oct 22:36 2014

sshfs's use of -x can stop X forwarding when ControlMaster is used


I just noticed that I did not have DISPLAY set in a login to a host for
which I configure "ForwardX11 yes". This turned out to be because I had an
sshfs mount as my first command in the current session to that host. sshfs
passes -x to ssh, which overrides my config. The connection is reused for
other logins (because I have ControlMaster auto configured for that host),
so I get no X forwarding.

I'm not sure what the best solution is, but how about passing "-o
ControlMaster=no' to ssh, so that it doesn't accidentally reconfigure a
session master?

I'm using sshfs 2.5 as packaged in Ubuntu 14.04.


Meet PCI DSS 3.0 Compliance Requirements with EventLog Analyzer
Achieve PCI DSS 3.0 Compliant Status with Out-of-the-box PCI DSS Reports
Are you Audit-Ready for PCI DSS 3.0 Compliance? Download White paper
Comply to PCI DSS 3.0 Requirement 10 and 11.5 with EventLog Analyzer
J. David Lee | 13 Sep 22:54 2014

Very odd problem with sshfs hanging


I'm doing some development in the Go language, and I'm trying to 
remotely mount a directory called "gocode" - this is a conventional name.

Running the command:

sshfs -odebug,sshfs_debug,loglevel=debug garcon:/home/johnl/gocode/ 

produces the following output:

SSHFS version 2.4
FUSE library version: 2.9.0
nullpath_ok: 0
nopath: 0
utime_omit_ok: 0
executing <ssh> <-x> <-a> <-oClearAllForwardings=yes> <-ologlevel=debug> 
<-2> <garcon> <-s> <sftp>

and hangs indefinitely. In fact I can't even get another terminal to 
open without dropping down the console as root and killing sshfs with -9.

The odd thing is that this behavior seems to depend on the name of the 
local folder. If I rename the local folder with

mv gocode xgocode

Then the mount works perfectly.

(Continue reading)

Daniel Malisani | 27 Aug 12:52 2014


Hi, I tired to probe several form to connect a remote folder wit write

The problem:
I can work correctly -without any problem- trough console- example: vim,
but when I open a file with X11 (kwrite, gedit, sublimetext, etc) I can't
write file.

the lastest command tried was:
Slashdot TV.  
Video for Nerds.  Stuff that matters.
Nick Edwards | 7 Jul 14:59 2014

sshfs reconnect problems


I'm having a problem with sshfs. I mount with:

sshfs nedward2 <at> /Volumes/HZZ/

The problem arises when my IP address changes, which happens frequently 
as I mainly work on a laptop with wireless connection. I've seen lots of 
other messages relating to sshfs hanging in this case, but my problem is 
somewhat different. As well as hanging, sshfs also attempts to reconnect 
automatically, even though I did *not* specify the "-o recconect" 
option. Unfortunately, my host does not support key based authentication 
or anything other than password authentication, so ssh authentication 
fails. The problem is that it keeps on trying, resulting in hundreds of 
failed ssh connection attempts from my username per hour, which causes 
alarm bells to ring for my sysadmin and my account being suspended. Is 
there some way to stop sshfs from attempting to reconnect? The hanging I 
can deal with, as I can just kill sshfs processes and reconnect by hand, 
but the problem is I don't always notice straight away when my IP 
address has changed, so I need to stop the automatic reconnects.

I recently upgraded my sshfs version, and I don't think I saw such 
problems with the old version.

 <at> c161219 ~$ sshfs --version
SSHFS version 2.5 (OSXFUSE SSHFS 2.5.0)
OSXFUSE library version: FUSE 2.7.3 / OSXFUSE 2.6.4

Many Thanks,

(Continue reading)

andrea biancalana | 2 Jul 18:50 2014

Re: pam_mount + sshfs

il giorno Mon, 16 Jun 2014 17:36:45 +0200  andrea biancalana <andrea.biancalana <at>> ha scritto:

> Hi,
> I'm trying to setup pam_mount + sshfs; I'd like to mount users home directories from a remote server.

It seems pmt-fd0ssh doesn't work fine: so I've put ssh="0" and password_stdin options.

I found known_hosts must be set system wide (/etc/ssh/ssh_known_hosts) and
now remote home-dir are mounted in /homeX mountpoint.

The problem is that it doesn't work if I put /home/%(USER) as local mountpoint.

Any idea?

Thanks, Andrea


<?xml version="1.0" encoding="utf-8" ?>

<debug enable="2" />

<volume fstype="fuse" path="/usr/bin/sshfs#%(USER) <at> xx.xx.xx.xx:" mountpoint="/homeX/%(USER)" 
options="password_stdin,reconnect,nonempty" ssh="0"/>
<logout wait="0" hup="0" term="0" kill="0" />
<mkmountpoint enable="1" remove="true" />
(Continue reading)