Bug#183659: cvs: fails to terminate its ssh child

[Colin Watson <cjwatson@debian.org>]

On Sat, May 24, 2003 at 01:33:41PM +0200, Cristian Ionescu-Idbohrn wrote:
> On Tue, 20 May 2003, Darren Tucker wrote:
> > In article <bab7fk$svc$1@spnode25.nerdc.ufl.edu>,
> > Ivan Krsul <ivan@krsul.org> wrote:
> > >Does anyone know of a way to get ssh to propagate the SIGTERM signal or to
> > >cleanup these orphan processes?
> >
> > This is an open bug.  There's a patch available (which should apply to
> > any recent version of OpenSSH).
> >
> > http://bugzilla.mindrot.org/show_bug.cgi?id=396
> 
> The bug report blames ssh (and so it should). I was able to reproduce the
> bug using ssh 1:3.6.1p2-2. Earlier ssh version most certainly suffer of it
> too.
> 
> The provided patch looks fairly simpel (5 lines added), but I did not test
> it. The ssh maintainer is hereby notified too.

Does using ssh >= 1:3.6.1p2-4 help?

openssh (1:3.6.1p2-4) unstable; urgency=low

  [...]
  * Backport upstream patch to pass monitor signals through to child
    (closes: #164797).

 -- Colin Watson <cjwatson@debian.org>  Sun, 27 Jul 2003 17:31:15 +0100

Cheers,

-- 
Colin Watson                                  [cjwatson@flatline.org.uk]