Problem with 10.7.3 (not crashing)

174 views
Skip to first unread message

andrei...@gmail.com

unread,
Mar 17, 2012, 6:14:08 AM3/17/12
to sshu...@googlegroups.com
Hello,

I am experiencing an annoying problem. First of all, I have got kernel panic only once - first time running sshuttle on 10.7.3, MBA with the latest firmware update.
Now I have the following problem -- after I stop sshuttle (ctrl+c twice), or loose connection, e.g. close the lid, I cannot restore it until I restart the laptop. The restarting takes quite a time, much longer than it would normally take.
Could you advice me how to restore sshuttle without restarting my laptop?

The command:
./sshuttle --dns -r and...@example.com 10.0.0.0/8 -vv

The last message I get trying to restore the connection:
firewall manager: starting transproxy.
 s:   Ready: 1 r=[4] w=[] x=[]
 s: <  channel=0 cmd=PING len=7
 s:  > channel=0 cmd=PONG len=7 (fullness=554)
 s: mux wrote: 15/15
 s: Waiting: 1 r=[4] w=[] x=[] (fullness=561/0)
>> ipfw -q add 12300 check-state ip from any to any
>> ipfw -q add 12300 skipto 12301 tcp from any to 127.0.0.0/8
>> ipfw -q add 12300 fwd 127.0.0.1,12300 tcp from any to 10.0.0.0/8 not ipttl 42 keep-state setup
>> ipfw -q add 12300 divert 12300 udp from any to 10.0.1.1/32 53 not ipttl 42
>> ipfw -q add 12300 divert 12300 udp from any 12300 to any not ipttl 42

The sshuttle version:
# git log
commit 274ee854d4f05fd55a7c1c94e0ffb29bed8008eb
Author: Avery Pennarun <apenwarr@**.*m>
Date:   Tue Feb 7 12:17:56 2012 -0500

    clean.do: don't forget to do version/clean.

Luis Gonzalez Sandoval

unread,
Mar 18, 2012, 2:52:00 PM3/18/12
to andrei...@gmail.com, sshu...@googlegroups.com
Did you tried flush the ipfw rules?

andrei...@gmail.com

unread,
Mar 19, 2012, 5:42:39 PM3/19/12
to sshu...@googlegroups.com, andrei...@gmail.com
Nope. How do I do that?

Andrei


On Sunday, March 18, 2012 7:52:00 PM UTC+1, Luis wrote:
Did you tried flush the ipfw rules?

Luis Gonzalez Sandoval

unread,
Mar 19, 2012, 5:48:05 PM3/19/12
to andrei...@gmail.com, sshu...@googlegroups.com
Try the following command:

sudo ipfw -q -f flush

andrei...@gmail.com

unread,
Mar 19, 2012, 5:51:08 PM3/19/12
to sshu...@googlegroups.com, andrei...@gmail.com
$ sudo ipfw -f flush

This doesn't help unfortunately. Any other way?

Andrei

andrei...@gmail.com

unread,
Mar 19, 2012, 5:58:35 PM3/19/12
to sshu...@googlegroups.com, andrei...@gmail.com
Thanks! Got my second kernel panic shortly after running it. :) But it didn't solve the problem.
I'll try with your fork to see if it helps.

Andrei


On Monday, March 19, 2012 10:48:05 PM UTC+1, Luis wrote:
Try the following command:

sudo ipfw -q -f flush

Luis Gonzalez Sandoval

unread,
Mar 19, 2012, 6:04:34 PM3/19/12
to andrei...@gmail.com, sshu...@googlegroups.com
I would recommend you to use that fork
is an older version but in my MacBook Pro runs stable.
 
I hope it helps.

Luis

andrei...@gmail.com

unread,
Mar 20, 2012, 3:03:45 AM3/20/12
to sshu...@googlegroups.com, andrei...@gmail.com
Still having the original problem even with the old code.

Andrei


On Monday, March 19, 2012 11:04:34 PM UTC+1, Luis wrote:
I would recommend you to use that fork
is an older version but in my MacBook Pro runs stable.
 
I hope it helps.

Luis

andrei...@gmail.com

unread,
Mar 20, 2012, 5:00:57 PM3/20/12
to sshu...@googlegroups.com, andrei...@gmail.com, apen...@gmail.com
There are five processes running as root, which I cannot kill:
root   14464    python ./main.py python -v -v --firewall 12296 12296
root   14396    python ./main.py python -v -v --firewall 12297 12297
root   14306    python ./main.py python -v -v --firewall 12298 12298
root    3678    python ./main.py python -v -v --firewall 12299 12299
root    2263    python ./main.py python -v -v --firewall 12300 12300

I have tried sudo kill -9 2263 without success. I am guessing that after killing them I will be able to run sshuttle again.
Is it correct and how do I kill them?

Andrei

Luis Gonzalez Sandoval

unread,
Mar 20, 2012, 5:09:07 PM3/20/12
to andrei...@gmail.com, sshu...@googlegroups.com, apen...@gmail.com

There are five processes running as root, which I cannot kill:
root   14464    python ./main.py python -v -v --firewall 12296 12296
root   14396    python ./main.py python -v -v --firewall 12297 12297
root   14306    python ./main.py python -v -v --firewall 12298 12298
root    3678    python ./main.py python -v -v --firewall 12299 12299
root    2263    python ./main.py python -v -v --firewall 12300 12300


which error message do you receive when you try to kill the process?

andrei...@gmail.com

unread,
Mar 26, 2012, 11:41:15 AM3/26/12
to sshu...@googlegroups.com, andrei...@gmail.com, apen...@gmail.com
Nothing. They just don't die. Is there any other way to tell them to quit?

Andrei

andrei...@gmail.com

unread,
Apr 10, 2012, 5:40:08 AM4/10/12
to sshu...@googlegroups.com
http://stackoverflow.com/questions/9794991/how-to-kill-unkillable-python-processes-running-as-root

I guess I need to add some kind of timeout in the source code, or listening for an exit signal. Any idea?

Andrei

klje...@gmail.com

unread,
Jul 17, 2012, 3:28:35 PM7/17/12
to sshu...@googlegroups.com
Any progress here?  Same issue on OS X Lion.
-Kyle

andrei...@gmail.com

unread,
Sep 3, 2012, 2:58:54 AM9/3/12
to sshu...@googlegroups.com
As of 9ce2fa0 the issue does not appear anymore.

- Andrei


On Saturday, March 17, 2012 11:14:08 AM UTC+1, andrei...@gmail.com wrote:

andrei...@gmail.com

unread,
Sep 3, 2012, 3:00:31 AM9/3/12
to sshu...@googlegroups.com, klje...@gmail.com
Yes, somehow resolved with 9ce2fa0 (or earlier commit).

- Andrei

Andrei Fokau

unread,
Sep 3, 2012, 7:14:42 AM9/3/12
to Kyle Jensen, sshu...@googlegroups.com
Hm, in addition to code update, I have also changed sshuttle parameters and run as:

./sshuttle  -r tun...@remote.host 10.20.30.0/24  -vv

i.e. no dns and more narrow ip range. I really hope it will help you as well, otherwise 
I will be horrified expecting this shit to come back.

Andrei


On Mon, Sep 3, 2012 at 1:03 PM, Kyle Jensen <klje...@gmail.com> wrote:
Andrei - I did not find that to be the case.  Still stuck with
unkillable processes in the "waiting" state.  I tried disabling the os
x firewall and also setting a global timeout in the socket module,
both to no avail.

-Kyle
Reply all
Reply to author
Forward
0 new messages