I am looking to build a machine to use for file backups and some light media serving (PhotoPrism, Calibre, and the like). My plan is to take a retired desktop, throw in every old drive I have lying around, and merge them with Greyhole.
I am a seasoned Linux user (in containers or on machines being managed by others). I am comfortable with CLI basics and should be able to use the system entirely over SSH and web UIs, so I am fine without a window manager. What I am not fine with is system administration (and do not want to spend time fixing issues that constantly pop-up).
Ideally, I would want to find a distro that I can install by accepting reasonable defaults, configure my storage and my various applications, create a cron job to periodically update packages and reboot the machine, and the machine will just keep working with no intervention on my part. I realize that that is an impossible goal, but I want to get as close as I can.
Maybe truenas.
Personally I just run proxmox at home with zfs for my array, then everything lives in a few vms / containers.
The reason that I was favoring Greyhole over ZFS is that I want to assemble a large, redundant storage volume out of a bunch of mismatched old disks and swap them out as they fill up or fail. I know it is very possible to do it with ZFS, but it seemed to not be the general use case and complicated.
Using unreliable hardware seems at odds with your “just want it to work” requirement.
Yes, but perfectly reliable hardware is impossible and every small gain in reliability drastically increases the price. Luckily, most hardware can be “fixed” by replacing the malfunctioning part. The only part that is not easy to replace is hard drives. For data disks, I have guides giving me step by step institutions on how to rebuild off of replicas. With the OS disk, I am depending on hopes and taking notes while installing.
Broken software is tricker to replace. I can uninstall and reinstall it, but I have to be careful to avoid catastrophic data loss. Also, broken software generally means a bad release, so I have to revert and periodically upgrade/revert to check if the issue has been resolved. And running old versions of one part of the system can cause incompatibility in other parts.
Your os disks should be raid 1. I’d recommend a pair of onboard nvme or cheap ssds for OS. That’s the part you don’t want to be fighting with if something does go wrong.
It’s trivial to replace a drive on zfs if one fails.
Hardware and software are both really reliable these days. If you’re running something like proxmox then you also get easy snapshots and rollbacks.