[381] in athena10
Re: Larvnet design discussion
daemon@ATHENA.MIT.EDU (Timothy G Abbott)
Tue Aug 5 15:59:44 2008
Date: Tue, 5 Aug 2008 15:58:58 -0400 (EDT)
From: Timothy G Abbott <tabbott@MIT.EDU>
To: Greg Hudson <ghudson@mit.edu>
cc: athena10@mit.edu
In-Reply-To: <1217965913.12433.112.camel@error-messages.mit.edu>
Message-ID: <alpine.DEB.1.10.0808051553590.21810@vinegar-pot.mit.edu>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; format=flowed; charset=US-ASCII
On Tue, 5 Aug 2008, Greg Hudson wrote:
>> Either way, I'd have the larvnet system check whether the machine was a
>> cluster machine before doing anything (to avoid annoying people with
>> strong privacy who use non-cluster workstations) and which thus can be
>> included in the default debathena-pam-config /etc/pam.d/gdm file.
>
> Current thinking is that all debathena-cluster machines are cluster
> machines, so there's no such thing as "check if this machine is a
> cluster machine". If you want a private workstation, you install
> debathena-workstation. If you really want a piece of cluster machine
> functionality like the tmp cleaner, you install that package by hand.
Right, I was just thinking it would be nice to not have two different pam
configuration packages if possible, but instead to use a single PAM
configuration and have the larvnet PAM module read /etc/athena/rc.conf (or
some similar file is included in debathena-cluster that programs like this
could use to determine whether it is a public workstation or not) to check
whether to nop. This is mostly an issue of managing complexity;
debathena-pam-config is already our most complex configuration package,
and it might be good to avoid making it any more complex by having it
either build two binary packages or having two versions of it.
-Tim Abbott
