build-the-nas §0 has no switch step, but port 15 moved from VLAN 30 to VLAN 40 #481

build-the-nas.md §0 is written as the complete list of what can be done before the drives land — BIOS, AMT, the static, the Kea reservation, the firewall rules, then the proof. It is missing a step that was evidently taken: the switch port had to move from VLAN 30 to VLAN 40.

Read off the switch on 2026-09-17:

• The static VLAN table gives CasaBonita 1,11-15, untagged 11-15 and ImaginationLAN 1,16-20.
• Port 15's PVID is 40, untagged only.
• smaug's MAC (OUI 4c:cc:6a) is learned on Port 15 in VLAN 40, and the port is up at 1000M full — the other three occupied ports on that segment are 100M.

Port 15 was ImaginationLAN's when the map was last read, on 2026-09-04, and it matched the documented map exactly on that day. So the move happened between then and now, and the obvious occasion is the install on 2026-09-16.

Why it is worth a step rather than a shrug. §0.3 and §0.4 are careful that the address is held twice, "because either alone is a single point of drift" — the same argument applies harder here. A NAS plugged into a port that is still on the lab segment gets a lab address and no rule reaches it, and the symptom is indistinguishable from a firewall fault, which is exactly the failure §0.5's note about the ports was written to prevent. The runbook also has no record of which port, so a successor re-cabling this machine has nothing to re-cable it to.

Suggested: a §0.4b — pick the port, move it into VLAN 40 untagged with PVID 40, and write the number down; plus a line in the §7 verification that the port's PVID is 40 and the MAC is learned there.

For the record, the wiki side is written: infrastructure/switching carries the new reading beside the 2026-09-04 one, and the physical diagram now draws the machine with the port on its edge.