~/Stuff(1) as well?

For a NAS, like, storage on the network, keep it as simple and as reliable as possible, so avoid Ubuntu and go to the core underlying OS: Debian.

Then just build up the functionality you need, is SMB, NFS, etc.

Personally, I went from OMV to a home built NAS, but went with Arch as that’s what I use elsewhere (btw), so am comfortable with it, but it’s bleeding edge which isn’t always the best if some functionality changes when you’re not ready for it.

If you’re going for a server running lots of containers, etc, then find whatever the container handler (docker?) is best on… I just put everything on bare metal, so can’t advise what’s best for containers… probably Debian again…

But, keep it simple.

I came from Nextcloud to syncthing, you’re in the right place.

… it’s a policy driven hell

Yeah, Point 1 here is exactly why I moved from Ubuntu to Arch ~10 years ago.

I was trying to get something working and found that the bug / feature had been fixed ~1 year earlier, but that version wasn’t in the repos… I couldn’t move forwards.

With Arch, all is well. And, I’m either reporting new bugs and helping to get things fixed, or I’m updating the wiki with any changes I notice.

To answer your first bit:

I went owncloud --> nextcloud --> syncthing + radicale.

Not looked back.

I run everything through a proxy in my home-built pfsense box.

You’re doing fine.

After seeing someone at work burnout, I’ll offer this advice:

Find what you enjoy doing and do nothing more (today). Itch only 1 scratch at a time.

As an analogy - consider you’ve moved into a newly built house and have an empty garden. No-one would expect you to create that perfectly first time around. Esp. in 1 weekend. It needs time to grow. Some things will need cutting down, some things will need moving. Animals will crap on it.

I think you’re trying to make it perfect, first time around. Perhaps as a fear of doing it “wrong”.

There is no wrong, it’s all a learning experience, doing things good enough for now and improving / breaking things later.

Ensure you know how to backup your files (3-2-1 rule) and the rest doesn’t matter.

I’ve re-written my ansible scripts a few times, but over months and years as I’ve learned what works best for my system.

For example, I had 1 complete script for each device. I can wipe the device (get it back on the network) and rebuild with no effort…

… then I realised that most of the scripts had very similar parts to tweak SSH and other settings, so then I learned how to call scripts from within scripts, which also meant using variables (facts) to work out if this is a 32b or 64b RasPi (for example)

That probably took 3 months

But I enjoy sitting in my garden and looking at it…

Longterm MythTv user here, watching the discussions

🍿

I want to search for a blog on this now…

Not arguing with you, it’s just a choice.

The question was whether Immich had to be executed from within a container system… and it doesn’t have to.

I guess that’s true.

I’m running it outside of a container and outside a VM… as there’s no abstraction layer on top of the underlying OS. Which I guess is inside the bare metal.

So, Yep.

I’m running it bare metal on my NAS.

No problems, plus I don’t have to do extra container stuff.

I’ve not seen that option, but I use syncthing instead of the phone application to sync my photos to a specific folder on my NAS which is then an external library for Immich.

TBH, I don’t want anything deleting anything automatically.

I’ll often delete newer pictures of temporary stuff but keep older pictures of my frinds & family, so, that’s not a feature I’d see any value in. It tends to just make me lazy and build up GBs of junk photos on my NAS (and backups…)

For Bitlocker? Good point, OP would probably need that for the new mobo’s TPM.

Using DHCP?

Windows stores static network configs in the registry, so with the new mobo’s NIC(s) if you try to set a static IP, Windows will complain that it already exists. Not a biggie as you’d just have to search the registry… if you’re using DHCP no problem.

Not quite clear there…

You’re copying data from the source, to harddrives… and then to a server with different drives?

Assuming it’s just lots of smallish data files / media and not OS files (ie don’t need symlinks, attributes, ownership, etc) then any backup software which generates hashes to be able to repair the archive during a restore would do.

Btrfs doesn’t need LVM, but I wouldn’t use that on mobile drives.

Or… is this one huge 80TB file?

And only if you get the torrent / magnet from the official site.

Is that headline number for the free or paid for version? I couldn’t tell from the article.

Looks like all downloads are direct from them, so it’s much harder for distros using torrents to know their install base, but I suspect they’re not much different.

I’ve seen increased activity with the Arch, Mint & Raspios torrents that I seed - although I don’t have cold figures to say how much by.

I think you’re looking for a calendar on a web page?

So, probably not what you meant, but Radicale is a really good caldav server I use for our calendars

It’s a server, you need clients (ie phones, etc) to see the calendars, but I found that no-one wanted a web calendar, they just used their phones… so maybe it’s an option…?

1st, definitely get backups offsite. Either cloud or drives at someone else’s home, but do that.

When (not if) something breaks you’ll need to fix it “now”

So, if you were intending on hosting a failover system in the cloud with Jellyfin, Adguard, Wireguard, etc. that won’t be a simple replica - you’ll need to redo your whole networking design.

IMHO, you’re better having physical spare parts / devices at home and focus on that.

If you’re running on an old PC, you’ll probably be better getting a newer, more efficient (lower electricity costs) - possibly smaller and quieter - device and moving stuff across… your old PC can then be the backup device.