oh I wasn’t talking about storage media. I’m talking about rack servers, switches, storage arrays (with new drives), etc., etc… The older hardware can wear out/break (I used to do MTTF/MIL-HDBK-217 calculations for avionics) but generally speaking it’s got a lot of life left in it by the time it hits the surplus market. It’s also usually designed with redundancies/failover mechanisms which means you don’t have to bodge together inferior solutions.

Gotta see some evidence on that claim. Older stuff is more power hungry no doubt about it, but especially old data centre equipment is waaay more reliable and built with some very nice creature comforts.

This.

Almost all of my gear is bought used: switches, server, even memory. My main server is an old Dell C6100 blade server I got for $250. My disk array is a 12-bay SAN that I found for $50 and took a chance on being able to get it working. It’s power hungry but it’s got redundant everything and I have spare parts on the shelf next to it.

I’ve been branching into ARM servers a little and right now I’ve got an RK3588 board with 32G of RAM. That’s new (and expensive for me) but I got a fibre channel array for $20 that I’m going to try to make work with it. $8 FC HBA and a $12 cable along with a $30 m2-to-PCIe adapter intended for eGPUs. I’m not going for speed here, but used data centre equipment is nice and some of it is dirt cheap because it’s too slow for “real” work.

Why not if (f & (F_1 | F_3)) {? I use this all the time in embedded code.

edit: never mind; you’re checking for both flags. I’d probably use (f & (F_1 | F_3)) == (F_1 | F_3) but that’s not much different than what you wrote.

It’s called the asshole tax. Don’t be an asshole and you won’t be charged.