Each device would always end up with the same IP… until I change any or all of them at the same time from one place.
What if you need to deploy a switch to a different network segment on which the statically assigned DHCP lease is invalid? Now you have an unreachable switch that may or may not be configured to handle the correct VLANs for that network segment.
If you move the device to a different subnet, you’re talking to a different DHCP server. Worst-case scenario: the switch would get a dynamic DHCP lease from the new server (which you could then make static and assign the desired address), and the switch’s web interface would be reachable immediately. Better still, you can copy the static DHCP definition from the old server to the new and adjust the IP, and the device would get the desired address upon initial power-up. This is how it works with any DHCP-capable device that I use (i.e., all of them except SwOS).
Unless you have a VERY large subnet allocated for management IPs, AND you setup a standard configuration on EVERY switch (i.e. port 1 is always trunk port, trunk port set to carry management VLAN, etc.), what you’re talking about is impractical.
Impractical? I do this now. SwOS is the only network device to which I have to assign an address manually before taking it into the field–in fact, at the moment it’s the only device of any type. Routers, switches, APs, IP cameras… everything else, I can grab a handful off the shelf and hit the road, and it doesn’t matter which one I connect where: it becomes reachable on the network, and I can configure it from there.
I’m not saying I don’t still have to configure switches differently for different locations; this is only about accessibility of the web interface. Just making the device reachable at its destination without special handling would save time. If SwOS had something analogous to MAC Winbox, it wouldn’t be so bad: you could still log on and configure it, even if its IP doesn’t match the network. But as it stands, if you don’t already know its subnet, the only ways to get onto a SwOS device are 1) a very lucky guess or 2) clearing its configuration and starting over. And when deploying more than one device (e.g., replacing a dozen RB750UPs with RB260GUPs), they aren’t interchangeable: you have to keep track of which one has the right address for which network.
Not only that, but you’re going to have to manually configure every one of those switches to set up your standard config, so where are you really saving time?
In this case, the only “standard config” I need would be to enable DHCP–which, I would argue, should be the factory default for networking equipment anyway. Like I said, this is mainly about making the SwOS web interface accessible on even a simple network without special handling.