

The daemon tracks file state, so the transfers start quicker because rsync doesn’t have to scan the filesystem.
The daemon tracks file state, so the transfers start quicker because rsync doesn’t have to scan the filesystem.
OWC and StarTech, I believe, have some Thunderbolt hubs.
It’s fine, treat it like a wear item.
Unfortunately, Optane and PCIe RAMdisks aren’t really things anymore. Those two would have been the best solution, but yeah, no.
A striped 0 array of NVME drives is the least bad option. Who really cares if it fails. It’s a cache; restart the job.
Parting out the laptops is probably more profitable than repairing them. There are people who will want to repair their equipment, and the supply of parts isn’t going to be infinite. If you have the space and time, holding onto parts would be worth it.
As for LibreBoot, you’re probably better off figuring out how to build your own boards around that. CoreBoot and LibreBoot are cool, but the equipment is old. People would want more modern equipment.
Off the top of my head, an Arm board with some MediaTek chips with LibreBoot which can pass Arm System Ready tests might be interesting. The SBC space is full of junk which isn’t upstreamed in Linux and thus can’t run a vanilla kernel, so there is an opportunity there. Something which could run Debian and OpenBSD would be the idea.
Not necessarily. Rsync deltas are very efficient, and not everything supports deltas.
It may very well be the correct tool for the job.
Anyway, problem fit wasn’t part of the question.