Mail Archives: cygwin/2004/03/18/13:58:18
On Thu, Mar 18, 2004 at 07:43:44PM +0100, Corinna Vinschen wrote:
> On Mar 18 10:29, Matt Berney wrote:
> > Thanks, I suspect that it is something in our environment that is causing this. Specifically, how do I use the strace tool in this context? Can I substitute the 'ssh' command with 'strace ssh' command?
>
> Ooh, no. The interesting part happens on the server side. You must
> start sshd under strace. Create a new service using cygrunsrv, e. g.
> called "trace_sshd", which runs `/bin/strace /usr/sbin/sshd -D'. Run
> it instead of the normal sshd service. The output will be automatically
> redirected to /var/log/trace_sshd.log. Be aware that the file might
> get *very* big. And it's a good idea to switch off privilege separation
> in your sshd_config file for this.
>
> Another step you can try first: There's a small chance that an Event
> Log entry has been created when the problem happens. I can't tell you
> how that might look like, but it's perhaps worth a try.
>
> Just musing: Is it possible that the /etc/passwd or /etc/group file
> gets frequently rewritten on the server machine? Perhaps something
> goes wrong there. It shouldn't but...
>
>
> Corinna
Here is another hypothesis. Cygwin gets the groups from a variety of
sources during setuid(). One of them is a call to NetUserGetGroups
to get the global groups from the logon server.
Failure of that call does not call a failure of setuid, because it
happens normally while running disconnected. So the problem could be
with your logon server or your LAN.
That hypothesis seems consistent with the outputs of your original
mail.
Fortunately there is a workaround: edit /etc/group and explicitly
include the user in question in the groups that should contain him.
Pierre
--
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Problem reports: http://cygwin.com/problems.html
Documentation: http://cygwin.com/docs.html
FAQ: http://cygwin.com/faq/
- Raw text -