Sounds like all you need is an Ext4 USB drive with a LUKS key on it. Then add kernel parameters like

rd.luks.key=UUID=/.keys/TheKey:LABEL=KEYS rd.luks.options=discard,keyfile-timeout=10s

in GRUB and it’ll autoboot.

Pull the key and power down and you’re back to normal. I use it in a low threat model environment so I can hit reboot and go get a coffee and come back to a DE.

ETA: sorry, got the timeout format wrong, I don’t use it.

Ah, so it is vibe coded [I created] was a trigger. Let people know, it might not actually be bad if you know coding already, but it’s good to let people know, not real maintainable otherwise.

Pretty much on the low side, but you’ve not been up long. Using key based login you’re fine.

Nah, the fun is learning form others mistakes. Thanks for a fun read :}

Seems pretty plausible, not 3-2-1 yet, but on the way, and should get the habits established well enough. Just having an offline backup is a huge step up from most. Consider a waterproof box (perhaps buried) in the back yard instead of just another room (in case of fire / flood).

If you have a friend with a similar setup, or who perhaps wants one, you can sync over internet and both get your offsite without the expense of online backups or the inconvenience of lugging HDDs around.

I run podman containers on my bazzite machines, basically you convert a docker-compose file to a .container file, here’s a bunch of examples, nextcloud is there, drop it in ~/.config/containers and run systemctl daemon-reload and it’s now a systemd unit that you start stop etc like any other. Updates are with podman autoupdate.

You can use podlet to convert docker-compose files (90% it works, otherwise it gets you 90% of the way there). It’s basically the fedora (/redhat) way to run containers.

I have no idea where you got it not being recommended (but adding to the main image sure is discouraged), and it’s certainly better than adding a vm for containers, which pretty much defeats the purpose of containers (to run using your main kernel, but contained).

I’ve been running my arr stack (with gluetun in a pod) etc this way for years now, very trouble free. Here’s a immich example.

It’s a bit of a learning curve, but it pays off.

Been using an quadlet podman arr stack for a year or two, pretty damn bulletproof once set up, easier to read, rootless, SELinux enabled, systemd controlled, update with podman auto-update. Worth the time to learn.

podlet can help you hit the ground running. It can create Quadlet files out of Podman commands or even (Docker) Compose files. 90% of the time it works every time ;}, but even the oopses get you most of the way there.

My arr stack is set up in a pod which means they all have their own gluetun network and come up as one, but you can just use Network=container:gluetun in container files.

Yeah, feel the same, curated RSS feeds also contribute significantly.

The amount of BS flying daily in the ‘media’ and with AI slop makes every day April Fool’s anyway. A constructed environment means the BS detector doesn’t have to be on 24/7, which becomes tiresome.

more people use their servers exclusively for personal entertainment than I expected.

Uh-huh, think of it like jigsaw puzzles…

That said, I prioritize ease of maintenance and simplicity, still wouldn’t expect my family to pick it up in any reasonable amount of time, nor have the motivation, more’s the pity.

I’ve moved to podman (quadlet) containers mostly, easy to read and edit, secure (mostly userspace), systemctl integration, autoupdate. I’ve done my distrohopping, fedora (in my case bazzite immutable) isn’t going anywhere, does everything I need. I run fairly lean, but have a bunch of stuff that can be spun up at a whim that I don’t use daily. It’s entertaining without being a burden, and useful stuff just happens.

Honestly, ssh and btop cover most of my monitoring needs, serious stuff gets a notify-send to my laptop. I’ve tried the web gui stuff and I don’t look at it enough to justify it, I’m not a sysad monitoring hundreds of computers, it’s just a hobby.

Pinchflat

Thanks, looks like it could work, bit more than I want, and it’ll need integrating, but it is what it is. I’ll see if it’s worth the pain to transition.

Of course there’s this…

MonkCanatella @sh.itjust.works English

I just have a somewhat annoying yt dlp script that runs on a schedule on my computer that downloads all my subscriptions. I tried everything else including pinchflat and it wasn’t really much of an improvement tbh.

TheLeadenSea
@sh.itjust.works
English

Could you share that script please? I’d like to set up something similar
    MonkCanatella
    @sh.itjust.works
    English

    Give me a bit, I’ll set up a gist with the details!
        TheLeadenSea
        @sh.itjust.works
        English

        Thanks!

followed by crickets. Such is life.

So, cool and all, but I’ve been using freshrss with libredirect and freetube for this for ages, and it’s great when freetube works, which with the cat and mouse with youtube is less often than I’d like (kudos to freetube, great project).

Honestly what I’d really like is something to download rss yt vid to a cache with yt-dlp (which breaks much less), grab the sponsorblock and pipe to something light and distraction free like mpv or vlc. Clear the cache after a day or whatever.

Anyone seen anything like this ?

For say a keypass db you don’t need even that, Just sshd gets you rsync on your computer with cron or systemd timer / service… Personally I just use an old version of Syncthing-Fork though, security implications for local network are minimal.

Sounds good.

Sensible, limiting scope and knowing your limits are wisdoms all too many lack. Having something that fills the text/sound/vidya needs and is easy to spin up will find uses, doesn’t need to do everything.

To the end of being easy to spin up, which is likely to attract other developers in time if that’s something you’d like, consider wrapping it up in a docker container. It’s not that hard, basically follow your own instructions in a special docker build format.

Some things to think about.

Even ZFS now let’s you add a new drive to an array, and the sweet spot for $/TB is ~16-20TB at the moment, so maybe think about 2 or 3x16TB and add more later (also less power).

Consider manufacturer recertified (not refurbished) server drives from serverpartdeals.com or your local equivalent, after all RAID is there to let you survive a disk failure, it’s treated me well, and lets you avoid SMR drives.

You can mix drives of different sizes if you use Unraid or roll your own with mergerfs+snapraid (+OpenMediaVault perhaps). I do the latter, it’s a bit of a setup, but has the advantage that drives are just drives and you can use the working ones while rebuilding the array and you can recover accidental deletions (for a while), which brings me to ‘RAID is not a backup’.

For true data safety you should have an offline backup (i.e. drives that live disconnected from your computer except during backup, safe from lightning, accidental deletion etc.) and eventually an offsite copy.

Personally I think the AI bought all our drives from WD is likely BS (seems lightly supported) to goose their product prices, so hopefully it’ll blow over, but prices seldom go down, inflation catches up. Sigh.

Yeah, I have “Analyze Video Files” on, doesn’t get me a list of substandard files though, just sends the arr after stuff it’s probably already not finding.

tdarr

Hadn’t seen the Property search in here before, might get me most of the way there. Got it around somewhere, might have to spin it back up. Maybe I can raid it’s database as well. Thanks.

So, unless I didn’t dive deep enough, Configarr / Trash guides is mostly about setting up quality profiles and media paths and so forth, something I long ago sorted out to my satisfaction.

What I guess I was after was something to find stuff that has fallen through the cracks, highlighting stuff that doesn’t meet my standards and seeing whether I care enough to go looking for upgrades.

Strangely there doesn’t seem to be a simple app to run ffprobe over your library and populate a database for querying video quality, maybe I’ll get around to knocking one out one day, but today is not that day.

I don’t want to balloon the project

Fair cop, and no I haven’t really dived into Configarr and the trash guides (although I vaguely remember coming across them), oh joy, another rabbit hole. I do try to keep a simple stack, and what I have has served me well for years. But thanks, no need to reinvent the wheel if that handles my use case.

Having smaller projects with specific scope that do something well and can be plugged together is always preferable to some sprawling monstrosity. Used to be called the Unix way (pipe sed into awk etc.) and could stand to be revisited today. Best of luck.

I had a quick look, I think I could find a use for it but what I’d most be interested in is a dry run spitting out a list of missing / low res / low bitrate / stereo (I much prefer 5.1+), perhaps old codec, etc. media. Like many I have my own standards for what needs to be how good and so forth.

Ideally I could edit said list and put it back in as an active search list (perhaps chunking and prioritizing as well and iterating the process). Seems like this is 90% of the way there, any chance of an enhancement ?

Bit reluctant to just let someone else’s code go ham on my media library without a me in the loop step.

And ideally the storage will be encrypted and have basic privacy assurances.

Why would you trust a company to encrypt for you when Cryptomator exists ?

Also, a couple of 4TB drives for cold backup (one offsite) avoids another subscription.