On Fri, May 28, 2010 at 12:32:33PM +0200, Dejan Muhamedagic wrote:
> On Fri, May 28, 2010 at 11:07:53AM +0200, Lars Ellenberg wrote:
> > On Thu, May 27, 2010 at 09:48:15PM -0600, Tim Serong wrote:
> > > On 5/27/2010 at 10:24 PM, Lars Ellenberg <[email protected]> 
> > > wrote: 
> > > > On Thu, May 27, 2010 at 12:46:14AM +0200, Matthias Ferdinand wrote: 
> > > > > --On Wednesday, May 26, 2010 12:00:02 -0600 
> > > > > [email protected] wrote: 
> > > > > >> OK, but it'd be still better/easier to just use ssh with public 
> > > > > >> key authentication. For telnet, there is a python plugin 
> > > > > >> ibmrsa-telnet which could be modified for iLO. 
> > > > > >  
> > > > > > DISPLAY=dummy SSH_ASKPASS=/bin/my_cat_passwd_file.sh ssh somewhere 
> > > > > > my_cat_passwd_file.sh: 
> > > > > ># !/bin/sh 
> > > > > > cat /etc/passwd_file 
> > > > > >  
> > > > > > /etc/passwd_file: 0600 root root containing your password ;-) 
> > > >  
> > > > > thank you for your hints. SSH_ASKPASS did not work for me (using 
> > > > > password 
> > > > > auth), ssh keeps prompting for the password. Apparently SSH_ASKPASS 
> > > > > is for 
> > > > > passphrases only. 
> > > >  
> > > > No.  But as long as ssh _does_ have a tty, it will ask for the password 
> > > > on the tty ;-) 
> > > > Only if it does not find a tty, it will use the askpass hook. 
> > > >  
> > > > > but as the script now does the job I think will just leave it at 
> > > > > that. 
> > > >  
> > > > It needs to work for you, that is what matters. 
> > > 
> > > If you still want to fiddle around with SSH_ASKPASS, it might help to
> > > redirect stdin from /dev/null...
> > 
> > Nope.
> > ssh explicitly opens /dev/tty.
> > So for it to not use the tty, it needs to have no tty ;-)
> > To get rid of a tty, you usually do setsid.
> 
> Isn't there a ssh option for this, i.e. don't allocate tty?

Well, yes.
That's one way I tested my claim that it works ;-)
ssh -T into an other box, causing me to not have a tty there,
and then doing the DISPLAY=dummy SSH_ASKPASS=script trick.
But it won't help for the described (and already solved) problem. 
Best solution is to use key based auth.

> Though this is really getting out of control ;-)


-- 
: Lars Ellenberg
: LINBIT | Your Way to High Availability
: DRBD/HA support and consulting http://www.linbit.com

DRBD® and LINBIT® are registered trademarks of LINBIT, Austria.
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to