[clug] Distributed Administration

[steve jenkin]

Michael Manning wrote on 24/9/09 6:46 PM:
> Howdy,
> 
> Two time short suggestions might be Kickstart/anacoda (Redhat based) and
> AutoYast (OpenSuse based) for auto pre-configured installs.
> 
> Then management, I am not sure on?

The best I've seen was at IP Australia with Damien Lederer using Solaris.
All software was distributed as 'packages'.
Solaris allows 'config' files for packages - variants done this way.
The package list of each of their 60+ systems would collected regularly
(nightly) so they could recreate a set of systems at a-point-in-time for
testing/troubleshooting.

They had a central server for admin.
It acted as package repository & for network booting (jumpstart) (and
syslog & SSH 'master' and probably more).

They probably had some simple SSH script to iterate an action across all
or a selection of hosts (DEV, TST, PRD, DB's, APP-srv, Web-srv).

These guys had a serial console server + a SAN.
They'd select a set of 'heads', configure the SAN and then from the
serial console be able to boot each machine from a jumpstart image.

The keys components were:
 - jumpstart
 - packages
 - storage separated from machines
 - lights out via serial console & management server

There must have been more I was unaware of...

Bob Edwards has talked about their setup, IIRC based around 'netboot'.
Perhaps he or Steve Hanley can point to what they've already posted on this.

What do people who run large clusters/GRID computing do??

HTH
s


> 
> Cheers,
> Michael Manning
> http://www.michaelmanning.org/
> http://twitter.com/mmanning13



-- 
Steve Jenkin, Info Tech, Systems and Design Specialist.
0412 786 915 (+61 412 786 915)
PO Box 48, Kippax ACT 2615, AUSTRALIA

sjenkin at canb.auug.org.au http://members.tip.net.au/~sjenkin