.: Device not configured after mounting sshfs

[Sean Davis]

I have installed fuse4x 0.8.13 and sshfs 2.3.0 (via homebrew). I
wanted to do a simple test and got the error in the subject line (name
of server changed). Any ideas?


NCI-01856172-ML:tmp sdavis$ mkdir s4
NCI-01856172-ML:tmp sdavis$ sshfs sed...@example.com:/data/sedavis /
tmp/s4 -o daemon_timeout=3600
NCI-01856172-ML:tmp sdavis$ cd s4
NCI-01856172-ML:s4 sdavis$ ls
ls: .: Device not configured

NCI-01856172-ML:s4 sdavis$ sshfs --version
SSHFS version 2.3.0
fuse4x library version: FUSE 2.8.6 / fuse4x 0.8.13
fuse: no mount point

Thanks for any help.

Sean

[Anatol Pomozov]

On Wed, Feb 1, 2012 at 3:11 PM, Sean Davis <sean...@gmail.com> wrote:

I have installed fuse4x 0.8.13 and sshfs 2.3.0 (via homebrew).  I
wanted to do a simple test and got the error in the subject line (name
of server changed).  Any ideas?


NCI-01856172-ML:tmp sdavis$ mkdir s4
NCI-01856172-ML:tmp sdavis$ 

NCI-01856172-ML:tmp sdavis$ cd s4
NCI-01856172-ML:s4 sdavis$ ls
ls: .: Device not configured

"Device not configured" message most likely means that fuse daemon process is not available (died). Please make sure that it presents:

"ps ax | grep sshfs"

But better if you run sshfs with debug enabled and see if there are any obvious problems with it. Add "-f -d" flags (foreground and debug) to sshfs.

sshfs -o daemon_timeout=3600 -f -d sed...@example.com:/data/sedavis /tmp/s4

[Sean Davis]

Forgot to attach a debug log:

NCI-01856172-ML:tmp sdavis$ sshfs -f -C -d sed...@example.com:/data/sedavis /tmp/s4 -o daemon_timeout=3600

FUSE library version: 2.8.6

nullpath_ok: 0

unique: 0, opcode: INIT (26), nodeid: 0, insize: 56

INIT: 7.16

flags=0x00000000

max_readahead=0x00100000

   INIT: 7.16

   flags=0x00000010

   max_readahead=0x00100000

   max_write=0x00010000

   unique: 0, success, outsize: 40

unique: 1, opcode: OPENDIR (27), nodeid: 1, insize: 48

   unique: 1, success, outsize: 32

unique: 0, opcode: ACCESS (34), nodeid: 1, insize: 48

   unique: 0, error: -78 (Function not implemented), outsize: 16

unique: 1, opcode: GETATTR (3), nodeid: 1, insize: 56

getattr /

   unique: 1, success, outsize: 136

unique: 0, opcode: STATFS (17), nodeid: 1, insize: 40

statfs /

   unique: 0, success, outsize: 96

unique: 1, opcode: STATFS (17), nodeid: 1, insize: 40

statfs /

   unique: 1, success, outsize: 96

unique: 0, opcode: STATFS (17), nodeid: 1, insize: 40

statfs /

   unique: 0, success, outsize: 96

unique: 1, opcode: STATFS (17), nodeid: 1, insize: 40

statfs /

   unique: 1, success, outsize: 96

unique: 0, opcode: STATFS (17), nodeid: 1, insize: 40

statfs /

   unique: 0, success, outsize: 96 

[Anatol Pomozov]

Hi,

Also show me your output for

$ mount -t fuse4x

I am interested to see if there are several fuse daemons try to map to the same directory.

[Sean Davis]

On Wednesday, February 1, 2012 6:45:31 PM UTC-5, Anatol Pomozov wrote:

Hi,

Also show me your output for

$ mount -t fuse4x

NCI-01856172-ML:s4 sdavis$ mount -t fuse4x

sed...@example.com:/data/sedavis on /private/tmp/s4 (fuse4x, nodev, nosuid, synchronous, mounted by sdavis)

Thanks for taking the time!

Sean

[Anatol Pomozov]

Hi

On Wed, Feb 1, 2012 at 4:38 PM, Sean Davis <sean...@gmail.com> wrote:

On Wednesday, February 1, 2012 6:45:31 PM UTC-5, Anatol Pomozov wrote:

Hi,

Also show me your output for

$ mount -t fuse4x

NCI-01856172-ML:s4 sdavis$ mount -t fuse4x

sed...@example.com:/data/sedavis on /private/tmp/s4 (fuse4x, nodev, nosuid, synchronous, mounted by sdavis)

This output looks fine, assuming that you have only one sshfs process running. Check it with "ps ax | grep sshfs".

sysctl information also might be helpful here. Please show me

$ sysctl -w vfs.generic.fuse4x.

How to repro your problem? Do you have this problem since you installed sshfs or it happened after some event (e.g. sshfs was running overnight and then stopped working)?

[Sean Davis]

On Wednesday, February 1, 2012 9:10:57 PM UTC-5, Anatol Pomozov wrote:

Hi

On Wed, Feb 1, 2012 at 4:38 PM, Sean Davis <sean...@gmail.com> wrote:

On Wednesday, February 1, 2012 6:45:31 PM UTC-5, Anatol Pomozov wrote:

Hi,

Also show me your output for

$ mount -t fuse4x

NCI-01856172-ML:s4 sdavis$ mount -t fuse4x

sed...@example.com:/data/sedavis on /private/tmp/s4 (fuse4x, nodev, nosuid, synchronous, mounted by sdavis)

This output looks fine, assuming that you have only one sshfs process running. Check it with "ps ax | grep sshfs".

sysctl information also might be helpful here. Please show me

$ sysctl -w vfs.generic.fuse4x.

NCI-01856172-ML:s4 sdavis$ sysctl vfs.generic.fuse4x

vfs.generic.fuse4x.control.kill: -1

vfs.generic.fuse4x.control.print_vnodes: -1

vfs.generic.fuse4x.counters.filehandle_reuse: 0

vfs.generic.fuse4x.counters.filehandle_upcalls: 9

vfs.generic.fuse4x.counters.lookup_cache_hits: 0

vfs.generic.fuse4x.counters.lookup_cache_misses: 1

vfs.generic.fuse4x.counters.lookup_cache_overrides: 0

vfs.generic.fuse4x.counters.memory_reallocs: 0

vfs.generic.fuse4x.resourceusage.filehandles: 1

vfs.generic.fuse4x.resourceusage.filehandles_zombies: 0

vfs.generic.fuse4x.resourceusage.ipc_iovs: 4

vfs.generic.fuse4x.resourceusage.ipc_tickets: 2

vfs.generic.fuse4x.resourceusage.mounts: 1

vfs.generic.fuse4x.resourceusage.vnodes: 1

vfs.generic.fuse4x.tunables.admin_group: 80

vfs.generic.fuse4x.tunables.allow_other: 0

vfs.generic.fuse4x.tunables.iov_credit: 16

vfs.generic.fuse4x.tunables.iov_permanent_bufsize: 524288

vfs.generic.fuse4x.tunables.max_freetickets: 1024

vfs.generic.fuse4x.tunables.max_tickets: 0

vfs.generic.fuse4x.tunables.userkernel_bufsize: 16777216

vfs.generic.fuse4x.version.api_major: 7

vfs.generic.fuse4x.version.api_minor: 16

vfs.generic.fuse4x.version.number: 0.8.13

vfs.generic.fuse4x.version.string: 0.8.13, Feb  1 2012, 16:53:13

NCI-01856172-ML:s4 sdavis$ ps ax | grep sshfs

37899 s003  S+     0:00.01 sshfs -o daemon_timeout=3600 -C -f -d sed...@example.com:/data/sedavis /tmp/s4

38004 s004  R+     0:00.00 grep sshfs

NCI-01856172-ML:s4 sdavis$ 

 

How to repro your problem? Do you have this problem since you installed sshfs or it happened after some event (e.g. sshfs was running overnight and then stopped working)?

 

This is a fresh install.  I have had macfuse on the system before, but I uninstalled a while back.  I can't see any evidence that this is an issue here, though.

Thanks again.

Sean

[Anatol Pomozov]

Hi,

"Device not configured" error means that kernel does not receive any response from fuse daemon (sshfs in your case). But from information you gave me everything looks fine.

BTW Does "ls" hangs or it returns the error instantly?

Ok, there are several thing you might want to try:
- check that ssh works fine with that url. Try "ssh sed...@example.com:/data/sedavis"
- check kernel logs, it is either /var/log/kernel.log or /var/log/system.log, do you see any thing related to fuse or sshfs?
- try to stop sshfs then unload fuse4x kernel extension. Make sure that "ls" does not return any error (at this point mount dir is just a regular directory). And then run sshfs again.
- try to reboot your computer

- try different flags for sshfs, e.g. sshfs us...@some-host.net:/home/user/ ~/mountpoint -oauto_cache,reconnect,defer_permissions,negative_vncache

[Andrew Berhow]

I am having, from what I can see, the exact same problem. Has there
been any resolution to this yet?

Thanks,
Andy

On Feb 2, 2:25 pm, Anatol Pomozov <anatol.pomo...@gmail.com> wrote:
> Hi,
>
> "Device not configured" error means that kernel does not receive any
> response from fuse daemon (sshfs in your case). But from information you
> gave me everything looks fine.
>
> BTW Does "ls" hangs or it returns the error instantly?
>
> Ok, there are several thing you might want to try:

> - check that ssh works fine with that url. Try "ssh seda...@example.com:

> /data/sedavis"
> - check kernel logs, it is either /var/log/kernel.log or
> /var/log/system.log, do you see any thing related to fuse or sshfs?
> - try to stop sshfs then unload fuse4x kernel extension. Make sure that
> "ls" does not return any error (at this point mount dir is just a regular
> directory). And then run sshfs again.
> - try to reboot your computer

> - try different flags for sshfs, e.g. sshfs u...@some-host.net:/home/user/

> ~/mountpoint -oauto_cache,reconnect,defer_permissions,negative_vncache
>
>
>
>
>
>
>
> On Wed, Feb 1, 2012 at 8:03 PM, Sean Davis <seand...@gmail.com> wrote:
>
> > On Wednesday, February 1, 2012 9:10:57 PM UTC-5, Anatol Pomozov wrote:
>
> >> Hi
>

> >> On Wed, Feb 1, 2012 at 4:38 PM, Sean Davis <seand...@gmail.com> wrote:
>
> >>> On Wednesday, February 1, 2012 6:45:31 PM UTC-5, Anatol Pomozov wrote:
>
> >>>> Hi,
>
> >>>> Also show me your output for
>
> >>>> $ mount -t fuse4x
>
> >>> NCI-01856172-ML:s4 sdavis$ mount -t fuse4x

> >>> seda...@example.com:/data/**sedavis on /private/tmp/s4 (fuse4x, nodev,

> > seda...@example.com:/data/sedavis /tmp/s4

> > 38004 s004  R+     0:00.00 grep sshfs
> > NCI-01856172-ML:s4 sdavis$
>
> >> How to repro your problem? Do you have this problem since you installed
> >> sshfs or it happened after some event (e.g. sshfs was running overnight and
> >> then stopped working)?
>
> > This is a fresh install.  I have had macfuse on the system before, but I
> > uninstalled a while back.  I can't see any evidence that this is an issue
> > here, though.
>
> > Thanks again.
>
> > Sean
>
> >>> Thanks for taking the time!
>
> >>> Sean
>
> >>>> I am interested to see if there are several fuse daemons try to map to
> >>>> the same directory.
>

> >>>> On Wed, Feb 1, 2012 at 3:31 PM, Sean Davis <seand...@gmail.com> wrote:
>
> >>>>> On Wednesday, February 1, 2012 6:11:50 PM UTC-5, Sean Davis wrote:
>
> >>>>>> I have installed fuse4x 0.8.13 and sshfs 2.3.0 (via homebrew).  I
> >>>>>> wanted to do a simple test and got the error in the subject line
> >>>>>> (name
> >>>>>> of server changed).  Any ideas?
>
> >>>>>> NCI-01856172-ML:tmp sdavis$ mkdir s4

> >>>>>> NCI-01856172-ML:tmp sdavis$ sshfs seda...@example.com:/data/**seda***
> >>>>>> *vis /

> >>>>>> tmp/s4 -o daemon_timeout=3600
> >>>>>> NCI-01856172-ML:tmp sdavis$ cd s4
> >>>>>> NCI-01856172-ML:s4 sdavis$ ls
> >>>>>> ls: .: Device not configured
>
> >>>>>> NCI-01856172-ML:s4 sdavis$ sshfs --version
> >>>>>> SSHFS version 2.3.0
> >>>>>> fuse4x library version: FUSE 2.8.6 / fuse4x 0.8.13
> >>>>>> fuse: no mount point
>
> >>>>> Forgot to attach a debug log:
>

> >>>>> NCI-01856172-ML:tmp sdavis$ sshfs -f -C -d seda...@example.com:/data/*
> >>>>> *seda**vis /tmp/s4 -o daemon_timeout=3600

[Anatol Pomozov]

Hi

On Thu, Feb 9, 2012 at 3:10 PM, Andrew Berhow <and...@berhow.com> wrote:

I am having, from what I can see, the exact same problem. Has there
been any resolution to this yet?

It is difficult to tell what is going on with fuse4x/sshfs without additional details. Ideally if you could provide a reproduce case - if I can reproduce it on my local computer then it is easier for me to diagnose the issue.

Could you please give me:

- Version of your operation system. "uname -a"

- Do you see anything related to fuse4x in /var/log/kernel.log & /var/log/system.log

Anything else that might be related to this issue. e.g. does it happen with all ssh server or just with some specific server.

[Anatol Pomozov]

Hi,

Just to wrap this discussion - the error says that the user-space
daemon (sshfs) does not exist. There are many reasons why it happens.
One obvious reason is that the process has crashed (e.g. segmentation
fault).

I think I saw a "segmentation fault" with sshfs when I had
incompatible version of libfuse and sshfs. sshfs was compiled against
production version of libfuse and run with debug version of libfuse.
It seems those versions have binary incompatibility so sshfs has
crashed. Once I compiled/linked sshfs against development version
everything works fine. Another reason is mixing libraries from
different fuse flavors (fuse4x, macfuse, osxfuse).

An advice for you is to uninstall fuse4x/macfuse, libfuse, sshfs and
reinstall it back. Make sure that you install it from the same source
- if you use binary packages, then use binary packages both for fuse4x
and sshfs. If you use package managers (macports, homebrew) install
both packages using the same manager.

I hope this helps.

[Erik Gilling]

I was having this exact same problem today.  I was trying to mount "sshfs username@host:local ~/local"   ~/local on the remote machine was a symlink to /usr/local/username.  If I instead mounted with "sshfs username@host:/usr/local/username ~/local" everything worked fine.

Cheers,

   Erik

[Anatol Pomozov]

Hi, Erik.

Mounting remote symlink should be fixed in the next fuse4x version
(0.10.0). You might want to install development build from here
http://dl.dropbox.com/u/3842605/Fuse4X-0.10.0-beta.dmg and make sure
that it works as expected.

Thanks for the report.

[Miguel Wahnon Monteiro]

my problem was a soft link problem. I was trying to reach a remote folder from a symbolic link instead of real path.

www -> /var/www

ssh me@host:/home/me/www 

instead of

ssh me@host:/var/www