Re: baffling ssh problem

[Wayne Topa <linuxone@intergate.com>]

Douglas A. Tutty(dtutty@porchlight.ca) is reported to have said:
> On Wed, Sep 19, 2007 at 12:54:57PM -0400, Wayne Topa wrote:
> > Douglas A. Tutty(dtutty@porchlight.ca) is reported to have said:
> > > On Tue, Sep 18, 2007 at 10:38:50PM -0400, Wayne Topa wrote:
> > >  
> > > > I've been messing with that for 2 weeks.  I must not have used a good
> > > > enough search phrase on Google.
> > > > 
> > >  
> > > Just for completeness, you didn't need a google search.  The info is
> > > somewhere around line 721 in the ssh(1) man page. :))
> > 
> > Not on my boxen Douglas.  It ends on line 579.  I just read the man
> > page, again, and still don't see it. I'm only half way through my
> > first cup of coffe so I will go back and read it again after cup 3-4. 
> > It must be in the AUTHENTICATION section.
> 
> Its in the FILES section, where it describes the purposes of each of the
> files needed for ssh to run.  It describes the permissions required on
> each.  This is on my Etch amd64 box, standard 80x25 screen.

All of the files in that section did have the correct permissions. It
was the home/USER directory that was at fault.  It was 770 when it
should have been 755 or 750.
> 
> > 
> > Another question popped into my head while trying to go to sleep last
> > night. 
> > 
> > Being that the 775 permission on the server home directory was the
> > problem, why was it only the AMD box that didn't work.  All 6 linux
> > boxes are running uptodate testing.  That was why I concentrated on
> > the AMD box as being the problem, and not the server.  ;-(
> 
> Did you perchance scp these directories from another box?  If not, how
> were they created?  Mine are generally created when I run ssh-keygen
> which is the first ssh-related thing I do on a box.
No, not scp, ssh-copy-id.  It creates the directory on connection to
the other box.  Yep, ssh-keygen creates the .ssh/id_dsa.pub on the client 
side.

I'm still trying to figure out if there is a bug in something on the
server or if there is one on the AMD64 box.  So far everything looks
OK.

Wayne

-- 
Real computer scientists only write specs for languages that might run
on future hardware.  Nobody trusts them to write specs for anything homo
sapiens will ever be able to fit on a single planet.
_______________________________________________________