open air for me too.
GPUs are zip tied to racks.
the room is the case and the windows are the exhaust.
Started my GPU farms like the above in 2016 and never looked back - ease of maintenance is the key.
Dust is the only pitfall though - I gave up on this long time ago and most of my 470/480 rigs still running at peak non-stop eventhough they are bloody dusty and grimy.
Well ..... as long as critical parts still works fine like fans for PSU, GPU and CPU - will just leave them alone. All of them are well ROIed several times over, so just leave them alone.
ease of maintenance you say?
how about real time (powered on and mining) pulling off slow spin/not spinning GPU fans and re oiling-plugging back in fans? hehe
no need to keep the area squeaky clean, for me i'm just making sure dust build up does not come close to possible "harmful" level to electronics----> i just come inside the room semi naked with face mask and a blower.. blowing the dust that accumulated in the electronics (all powered on and mining)
Most of the Gigabyte RX470s fans died premature death - I had to replace several dozens of them. The Sapphire ones are very good and lasts longer - most still running since 2016. The motherboards (circa 2016/17 ASrock Pro BTC) are beyond recognition - the dust and grime probably evolved into a protection layer for the electronics thus still running until today. I should take a pics next time - we should start a most "dirtiest GPU warrior rig" pix contest.
In all the years that I was mining. I've never had a motherboard, CPU, or RAM ever fail on me.
The parts which have the highest failure rate are the GPU fans, then maybe the GPUs itself, then the Risers, then the PSUs, then the harddrives.
The most reliable GPUs I've had were the cheap and loud Reference GPUs from all generations. Don't think I've ever had one fail.
The most reliable PSU was probably the cheap $40 Corsair 750CW. Had tons of those and suprisingly it never failed while my Gold rated 1600W did fail. Ironic.
