Finder locks up requiring hard reboot during use of sshfs

[GoogleCodeExporter]

What steps will reproduce the problem?
1. I was mounting a filesystem using sshfs, over a fast local network, and 
using exclusively
 with Xemacs. After some hours of use file saves in Xemacs failed, and finder hung, requiring a 
hard reboot.

Currently using OS X 10.4.8, MacFuse Core 0.2.1, sshfs 0.1.0

syslog has:

Feb 24 16:06:34 frenzal kernel[0]: MacFUSE: OUCH! daemon did not give fh 
(type=1, err=4)

at around the time of the problem, which may be related.

This is reasonably vauge bug report, but maybe someone else sees something 
similar and can 
add more information.

I will try and compile from source, and add some extra debugging to try and 
isolate the problem.

Original issue reported on code.google.com by ben.les...@gmail.com on 24 Feb 2007 at 12:38

[GoogleCodeExporter]

It's possible that your ssh connection might just have timed out for some 
reason. The "daemon did not give fh" 
message in your case has error # 4, which means EINTR (interrupted system call).

Original comment by si...@gmail.com on 24 Feb 2007 at 6:36

[GoogleCodeExporter]

By the way, even though the Finder might not be responding, try killing (kill 
-9) the sshfs process from the 
command line.

Original comment by si...@gmail.com on 25 Feb 2007 at 12:02

[GoogleCodeExporter]

Please also look at issue 113, which might be related.

Original comment by si...@gmail.com on 1 Mar 2007 at 12:41

[GoogleCodeExporter]

This does indeed look like 113. In particular I know that emacs uses renames a 
lot during its file saving. I am 
now trying with the single thread option and also debug turned on, so hopefully 
I will be able to get a better idea 
of what is happening.

Original comment by ben.les...@gmail.com on 1 Mar 2007 at 2:53

[GoogleCodeExporter]

Can you try the updates to MacFUSE (0.2.5) and SSHFS (0.2.0) and see if you 
still have this issue?

Original comment by si...@gmail.com on 19 Apr 2007 at 9:45

[GoogleCodeExporter]

Sorry, haven't had a chance to look at this lately. I've just downloaded 0.3 (of
MacFUSE and SSHFS). I'll let you know how it goes.

Original comment by ben.les...@gmail.com on 16 May 2007 at 5:24

[GoogleCodeExporter]

I experienced the same initial symptoms I was originally seeing. (E.g: Xemacs
complaining about an interrupted system call.) Now I eventually get a finder 
dialog
complaining the the remote filesystem can't be accessed, and gives me the 
options of
force unmounting it. After this occurs I can reconnect with no problems at all.

This time I don't see any errors in the system log.

Original comment by ben.les...@gmail.com on 16 May 2007 at 5:52

[GoogleCodeExporter]

ben.leslie: Do you just see this happen with xemacs or even otherwise?

In the case of xemacs, it would be very interesting to see:

1. If the problem occurs when you do not use X (that is, use "xemacs -nw").

2. If the problem occurs with Apple's bundled emacs.

Original comment by si...@gmail.com on 17 May 2007 at 6:36

[GoogleCodeExporter]

ben.leslie: by any chance, do you use Little Snitch?

<http://code.google.com/p/macfusion/issues/detail?id=111#c3> was originally 
focused on ftp, then expanded 
to include other possibilities, after which the original reporter found a 
workaround (and hopefully their solution).

(Excuse my interruption from the MacFusion neighbourhood. I'm particularly 
interested in _anything_ that is a 
peculiar problem for either community.)

Original comment by grahampe...@gmail.com on 17 May 2007 at 8:02

[GoogleCodeExporter]

asingh: Currently I really only use this for xemacs. I'll try with both xemacs 
-nw, and emacs too.

grahamperrin: Not using Little Snitch.

Currently when I get the interrupted syscall error, I'm able to unmount, then 
remount and this is a reasonable 
workaround for now. (Still better than using emacs TRAMP).

Original comment by ben.les...@gmail.com on 30 May 2007 at 7:57

[GoogleCodeExporter]

After yet more testing, I can confirm this seems to be Xemacs only problem. The 
emacs that comes with OS X 
never sees this problem. Current behaviour seems a little better. If I get the 
"interrupted syscall", stop for a few 
seconds, and tehn try and save, things seem to work.

Original comment by ben.les...@gmail.com on 17 Jun 2007 at 7:24

[GoogleCodeExporter]

Marking this as duplicate--more context in issue 214.

Original comment by si...@gmail.com on 26 Jun 2007 at 8:28

• Changed state: Duplicate

[GoogleCodeExporter]

Agree.

Original comment by ben.les...@gmail.com on 21 Sep 2007 at 6:59