Hey all, I’ve been doing a bunch of research on selfhosting the last few weeks as I’d love to lean on more open source projects for my daily productivity & entertainment. My main goal is to backup all my personal documents, photos, and videos (around 1tb so far over ~5 years, so not too demanding) and host a few services to access files on local storage (Immich, Jellyfin) and personal (paperless-ngx, homeassistant, morss). Although I’m not afraid to mess around learning Docker, I’d like to prioritize low maintenance in balance with relatively low long-term cost so that I don’t run into an issue that takes more than a day to restore access to my files/backups. I’d rather save that time for the fun stuff, like endlessly configuring HA automations.
All that said, I figure a decent solution would be to run a local NAS in RAID 6 with a cold storage HDD to swap whenever I transfer a bunch of files from my camera for local backup, and a remote backup at either my parents’ home or maybe eventually on another friend’s NAS. The main thing I’m wondering right now is if a prebuilt NAS (Synology, Asustor, etc.) is worth it in comparison to a custom built system for simple maintenance, reliable and low-bandwidth remote backup and recovery, and solid file sharing options for friends and family? I’ve heard SFTPGo is a great project for file transfers if going custom built, so I’m not completely worried about the last point, but it’d still be a nice bonus to not have to worry about another service.
My greatest fear is having to explain to my parents what a terminal is, so I’d like something reliable with a good price which I can hopefully maintain without crossing that bridge. I know most prebuilt NAS systems aren’t as cost effective or flexible for hosting a bunch of services also, so if I did go with a prebuilt, I would probably pick up a micro PC like a NUC or an old Dell Optiplex to network with the NAS for Immich, and maybe use some internal storage to keep some movies to stream with Jellyfin (unless there’s a limitation I’m not considering). Any advice?
single point of failure
That only makes sense when you’re talking about adding redundancy imo, because multiple devices also add more sources of failure. Personally I’d rather have everything failing all at once every 20 years (with backups ofc) than something different breaking all the time.