Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 28 Feb 2002 09:35:37 +0100
From:      "Cornelius, Peter" <peter.cornelius@comsoft.de>
To:        'Danny Howard' <dannyman@toldme.com>, "Thomas P. Holmes [ Systems ]" <tholmes@thebiz.net>
Cc:        Koroush Saraf <koroush.saraf@lmco.com>, freebsd-cluster@freebsd.org
Subject:   AW: Mass Upgrade and Maintenance questions
Message-ID:  <905777482C38D5119D0D00D0B7A06E2D0C7BC1@cspop.comsoft.de>

next in thread | raw e-mail | index | archive | help
High folks...

...many machines, identical users on all of them, same subnet, =
possibly the
same set of programs, etc. How come that, to me, this appears a prime
entirely-diskless/NIS candidate?

If so, check /usr/ports/net/etherboot, /etc/rc.diskless[12] and the =
handbook
(http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/diskless.html=
,
http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/nis.html). =
There
also are an article on diskless
(http://www.freebsd.org/doc/en_US.ISO8859-1/articles/diskless-x/article.=
html
) and some tricky examples in /usr/share/examples/diskless which took =
me a
little while to get the drift of; mount_null(8) and mount_mfs(8) are =
your
friends. Searching for 'diskless' on the web site also gives a lot of
results
(http://www.FreeBSD.org/cgi/search.cgi?words=3Ddiskless&max=3D250&source=
=3Dwww).

Dunno wether this is any help to you.

best regards,

Peter.

> -----Urspr=FCngliche Nachricht-----
> Von: Danny Howard [mailto:dannyman@toldme.com]
> Gesendet: Donnerstag, 28. Februar 2002 09:11
> An: Thomas P. Holmes [ Systems ]
> Cc: Koroush Saraf; freebsd-cluster@freebsd.org
> Betreff: Re: Mass Upgrade and Maintenance questions
>=20
>=20
> On Wed, Feb 27, 2002 at 10:45:00PM -0500, Thomas P. Holmes [=20
> Systems ] wrote:
> > > NFS?  I'd mount 4.5-RELEASE /usr/src on all these machines, make
> > > buildworld and buildkernel, then write a script that goes=20
> to each in
> > > turn installs both.  Maybe put the procedure in an=20
> install.sh and do
> > > like this:
> >=20
> > Are you saying to buildworld on each seperate machine with=20
> having just
> > the src code nfs mounted?  Or build on 1 machine and then=20
> installworld
> > on each machine?  If it is the latter, I've had problems the 2 =
times
> > I've tried that.  It could be differences in the machines I=20
> guess.  It
> > always seems to bomb with something like "command install not =
found"
> > halfway through.
>=20
> Sh!t happens.  Measure twice, cut once.  I'd have a couple of boxes
> reserved for guinea pig testing, myself.  I've found that =
installworld
> from multiple machines over NFS works, but there's danger any time
> you're upgrading the OS.
>=20
> > In our production work environment we have a set of master=20
> servers and
> > everything else are diskless clients of the masterservers.  =
Services
> > are "sandboxed" so that too makes life easier :)
>=20
> Yes, network booting is da bomb, and that is how I'd prefer=20
> to maintain
> 34 workstations. :)
>=20
> -danny
>=20
> To Unsubscribe: send mail to majordomo@FreeBSD.org
> with "unsubscribe freebsd-cluster" in the body of the message
>=20

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-cluster" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?905777482C38D5119D0D00D0B7A06E2D0C7BC1>