Apologies if this is the wrong section, can anybody point me in the right direction as to where to start with this exercise. I'm sure IPMI will enable the process but I need some guidance as to how to get an initial system booted.


In broad terms what steps would you take to automatically install, configure and manage Linux on an initial 20 servers from bare metal?

The servers are in a co-located data-center, not cloud based.

It can be assumed that the administrator has:

1. A workstation to administer the builds.

2. Command line access to the physical servers (either physical console or IPMI).

3. Access to a public/private key pair.

4. A choice of .deb or .rpm based distro.

5. Access to network infrastructure to support any build processes.

6. A spec for each server that requires a base build (all) and specific software (select machines).

7. All software to be installed, made available as native packages.

Thanks for your help in advance

ML

I would look at pxe, and tftp

Tftp often comes in as 'network boot' options in the bios.
The idea is the box loads a kernel by what is basically extensions of dhcp, and boots it; it then looks for an OS and installs it. Slackware has a good document on this explaining setup required, and it's in the usb-and-pxe-installers/ directory.