Ideas, use cases, and improvements - what are you doing with PICOTTY?

[morpheuslord]

PICOTTY started as a way to get a console into half-dead Proxmox nodes without walking to the rack. The core idea (a cheap Pico per machine, wired Ethernet, one hub, a browser dashboard) can go further than my homelab, and I want this thread to be where that happens.

Things worth posting here:

• Feature requests. If something in the firmware, hub, or dashboard gets in your way, or you keep wishing a capability existed, describe it. Rough ideas are fine. If it turns into something concrete, we can promote it to an issue.
• Your setup. How many nodes, what hardware, what you manage with it. Photos welcome. Unusual targets are especially interesting: routers, SBCs, lab equipment, anything with a serial console or USB port.
• Ways to do more with less. The point of this project is that a node costs a few dollars. If you found cheaper parts, other boards that work, enclosure designs, or wiring tricks, post them.

Resources. Guides, scripts, integrations with monitoring or automation tools, anything that helps people get running faster.

No format requirements. Half-formed ideas and "would this work?" questions are as useful as finished writeups. If you built something on top of PICOTTY, even a small script, I want to see it.

[jakobi]

I skimmed the docs and quite liked the level of detail and included reasoning.

However I wonder about the reason for choosing ethernet over wlan and the seeming lack(?) of not considering or ignoring the pico W even in the roadmap or some to be added portability section?

And similar to my question about the rm2 of the pico W, didn't wiznet make some updated ethernet hats and aren't there some pico devboards already with the wiznet module included? Even just a section with a raw sketch on likely portability issues to these would nicely round out the docs.

thx for your thoughts on these points & congrats for having just gotten CNX'ed :)
Peter.

[morpheuslord]

Yes, and agreed. On WLAN first: I did consider it, and my main pain point is the flakiness of the setup. The whole system is only as reliable as the network link, and that's the reason I went with Ethernet. That said, I am working on a Pico W version (ordered one, waiting for delivery). The plan is that it can either be the primary connection on its own, or act as a failover when both Ethernet and Wi-Fi are configured. The open question is isolation: for the management VLAN to work over Wi-Fi, the switch and the AP need the same VLAN access, or the deployment has to sit outside a VLAN entirely. Not sure yet which is better.

On the WIZnet boards: fair point, and I should cover them in the docs. I actually looked at the EVB-Pico boards early on. The reason I didn't go with them was stock. Every single store I called had none and was waiting on a restock, so I built around the W5100S HATs I could actually buy. For reference, the lineup: the W5500 HAT is a straight upgrade over the W5100S (more sockets, faster SPI), and there are the integrated boards, W5100S-EVB-Pico, W5500-EVB-Pico, and the W55RP20-EVB-Pico, which packs the RP2040 and W5500 into one chip. For the W5100S and W5500 boards, porting should be minor. CircuitPython's WIZNET5K driver handles both chips with the same API, so it's mostly pin assignments. The W55RP20 runs the SPI link over PIO, so I need to confirm library support before I claim it works. The W6100 (IPv6) is further out for me. I'll add a portability section to the docs with a compatibility sketch for these; good suggestion.

And thanks re CNX :)