Hey all,

So I have just been approached by a company who want to interview me for a good role but one of the criteria is that I have knowledge in building a new Linux server and then creating a template for all future builds??

If these were virtual I would understand but I am kind of stuck on what would be the best practice for this. How do you create a template on bare metal builds in corporations these days? Not done that before.

I have read about REAR (Relax and Recover) but I dont think that is used in mainstream businesses? Or is it?

I would appreciate your comments on this so I know what study for.

Thanks all

There are bare metal provisioning tools like Cobbler or Crowbar. Your interviewer may also thing of systems like Ansible.

I'm never laid hands on a server, but duckduckgo suggested pages like
https://www.linuxtechtips.com/2013/1...-in-linux.html
https://www.ibm.com/support/knowledg...te_server.html
so I suspect the answer is out there.

There is little difference between a virtual and a 'real' server in this context, aside from getting the OS spun up. Sure, a virtual server can be a blank already-installed OS that you need to copy over/activate/tweak, but aside from doing a manual load from PXE/DVD/USB/Whatever to the 'real' server, the differences end there. You begin with a configured Linux server.

I'd ask what these Linux servers are going to be DOING, rather than them just being servers. I typically will recommend building RPM's of custom-deployment things, and using an internal repository. Need a postfix box? Here's the RPM we created internally (and *KEEP UP TO DATE*), that has the configs, databases, and a set of firewall rules. Same for a web server, DHCP, or whatever else. Once you get the internal repository centralized, the rest is easy. If you need to change your active postfix config for a new relay host lets say, you change and test the config file on ONE box, and once it works, RPM-ize it, and shove it into the repository. Your other servers will pick it up, and apply it, along with a service restart. Let the package managers do the heavy lifting.

This also lets you control things, since using something like svn or git to push files to the repository gives you a trail, and lets you do version control. However, every company is different...I'd start by asking what infrastructure they have in place already to handle such things, and if they're satisfied with it.