(25-08-2020, 08:19 AM)z3bra Wrote:

(24-08-2020, 08:37 PM)freem Wrote: redundancy of my lab VMs: it implies doubling a full pxe infra, maybe port the fallback infra to systemd

How is systemd related to pxe ? Or maybe you're aggregating two topics here ?
Which hypervisor do you use, and what PXE system ?

systemd is not related, it's just to add some complexity and gain some experience with something I dislike. Honestly, it's not a top priority.
My VMs run under virtualbox, because it's the simplest one to use.
PXE is not really "one system", it's a combination of softwares for which there are several alternatives. The time I've spent on deploying those was mostly spent on configuring servers.
Basically, it is built from a dhcpd, a tftpd and an "nfsd". Add some pxelinux configuration there, a spoon of chroot, put a good dose of shell scripting, and if you know your distro, you should almost have it.
I tried it at 1st for work, but I was *really* unsatisfied by the result, debian does not really let itself run as diskless. So I've redone the stuff for myself, deploying a small set of VMs. The goal: to have a virtual network that allows me to deploy a very heavily customized debian. I could almost say a fork, despite the fact I don't (yet) rebuild packages there. I need the arch to be efficient 1st :)

Well, I failed my planning, being jobless does not help to hold those... I have time, but motivation varies too much.