so-called-hang-on-exit

[Eric Garff]

This isn't an actual reply to any one person, but here is my experience:

I have tested both the patches by Jani Jaakkol and Frank Cusack (Frank's
I tried both the pty and non-pty changes).  I still get the hang, here
is my setup.

This is entered as a cronjob on one server to run every 15 minutes:

/home/httpd/bin/cache/glb/gTest >>/home/httpd/bin/cache/glb/gTest.log
2>>/home/httpd/bin/cache/glb/gTest.log

  here are the guts of the script:
	/usr/local/bin/ssh hostname /bin/ls /some/dir | /usr/bin/wc -l ;
/bin/ls /some/other/dir | /usr/bin/wc -l ; /bin/ls /some/other/dir |
/usr/bin/wc -l"

The script is actually a php script that does a number of other things
before it executes the ssh (database to grab hostnames, etc.), but that
is how the ssh portion pans out.  After executing the command it then
starts a 20 second timer, and if the ssh session is still alive, it
kills it.  Also, if I run the script by hand (not crond) it completes
just fine (no hang).

Any ideas?  Your time is MUCH appreciated.

On Tue, 2002-08-06 at 12:27, Craig J Copi wrote:
> Nicolas.Williams at ubsw.com writes:
> >
> >I meant, what is the difference in behaviour in technical terms.
> >
> >E.g., "the BSD pty master gets EOF as soon as all processes in the pty's
> >controlling process group exit, whereas on Linux the pty master EOFs
> >only when there are no open file descriptors for the pty."
> >
> >And what server-side, user-level heuristic could "correct" for this
> >difference?
> 
> I believe that is the question nobody has answered.  There have
> been calls for someone to look into the guts (presumably kernel) and
> figure out what is going on.
> 
> Craig
> _______________________________________________
> openssh-unix-dev at mindrot.org mailing list
> http://www.mindrot.org/mailman/listinfo/openssh-unix-dev
-- 
Eric Garff
System Administrator
egarff at omniture.com
801.722.0134

Omniture
http://www.omniture.com
"Unique Questions, Precise Answers"