docs(install-deploy): reach a maintenance node with no DHCP server

[Bugs5382]

What and why

A maintenance node now gives itself an RFC 3927 link-local address in 169.254.0.0/16 when no DHCP server answers, and its console shows that address (CryptOS-PKI/cryptos-node#385, merged). So the boot and maintenance page no longer has to say there is no fallback.

• Prepare the machine: the network no longer needs DHCP; the old "no fallback yet" note is gone.
• What happens at boot: says what the node does when DHCP gets no answer, and that the Address line then shows the 169.254.x.x address.
• Plug a laptop in directly: a new section: give the laptop a 169.254.x.x address (macOS and Windows do it on their own; on Linux, NetworkManager's link-local method or ip addr add), read the node's address off the console, check the fingerprint, then cryptosctl --insecure --endpoint 169.254.x.x:443 status.

Refs CryptOS-PKI/cryptos-node#381

Verification

• The page describes what is on cryptos-node main now: the fallback in internal/init/linklocal.go and console.ReachableAddrs.
• Addresses are 169.254.x.x placeholders and RFC 5737 examples only.

How this was verified

CI is the proof for this change: the docs build checks the page and its links. The cryptosctl command needs a booted maintenance node; the appliance QEMU suite proves it end to end in a follow-up PR on cryptos-appliance.