[206] in athena10
Re: "mkserv public / mkserv clean"
daemon@ATHENA.MIT.EDU (Mitchell E Berger)
Wed May 14 16:37:47 2008
Message-Id: <200805142037.m4EKb1nf008116@byte-me.mit.edu>
To: Jonathan Reed <jdreed@MIT.EDU>
cc: athena10@MIT.EDU
In-Reply-To: Your message of "Wed, 14 May 2008 16:27:03 EDT."
<0B54C386-72BD-4862-BBD7-2D0113064B96@mit.edu>
Date: Wed, 14 May 2008 16:37:01 -0400
From: Mitchell E Berger <mitchb@MIT.EDU>
I think I may have slightly misled Jon. I'd have to think harder to
be sure (or read code more closely). I think mkserv public may in
fact punt things from /var/server right away (it should run the service.del
scripts). However, the full integrity check won't happen until a reboot,
which I think is what he was worried about. If that's correct, then
it's not much different from a flag file examined at boot time.
Mitch
> Discussion on zephyr points out that I'm completely on crack and
> mkserv public does not actually immediately cause removal of anything
> or verification. I'm not sure what I was actually thinking of. So
> "never mind".
>
> -Jon
>
> On May 14, 2008, at 4:15 PM, Jonathan Reed wrote:
>
> > I realize that mkserv is going away in Athena 10. I wonder,
> > however, if we'll have any mechanism by which users can turn a
> > workstation into a vanilla cluster configuration. OLC does find
> > itself in the position of telling users to use this command once in
> > a while (mkserv public is more common than mkserv clean). I realize
> > we plan to have a "public athena" flag file somewhere, but obviously
> > creating that file will not instantaneously cause the machine to fix
> > itself. Will there be a method by which a user can request that a
> > machine immediately revert to a clean state?
>
