Ebby

I had an Ubuntu system ate itself Christmas Eve at 1am. I think an update pooched some folder permissions and wiped the firewall settings.

It started with “Huh, the network drives won’t mount”, progressed to containers failing to start with “hardware not found” GPU-related errors -oh yeah, I still have to check that and make sure it is working- and yadda yadda… 13 hours later I got services back up and running.

Oof, today I slept in until 10. Ahhhhh…

Man, I wanted my Zoraxy migration to go smoothly, but it’s been stubborn as hell for me. 80% of my services transfered well, but a couple docker containers don’t like it and never got letsencrypt wildcard dns up and running; always an error. Still working on that in my spare time.

Yahoo! Congrats to the Immich teams and developers!

Bingo. I’ve used Synology for ages and while they dont last forever, I get a lot of use out of them and re-buy them usually with an upgrade in mind.

But the new hard drive policy broke that cycle. I don’t put up with that. I replaced one with a UGreen NAS last month. It’s too early to tell how I feel about it. Docker is there and containers spin up pretty easily. Rumor has it there is hardware support for video encoding too, though I haven’t gotten around to testing it.

Even though there are some cloud services like remote server management, proxies, and 3rd party integration, I do actually have to run the software myself on my hardware. Hence, self hosted.

I have the same sensors, but on many windows. If I have windows on either side of the house open simultaneously and there is a favorable temp difference outside vs inside, an automation turns on an air exchange fan. If they are closed, I use advanced heating control

I get what you are saying, but in the case of the internet, you need an IP address to connect rather than simply exist with a computer. Someone needs to know where to send the data.

There are however free connections: unsecured neighbors wifi, city wifi, hotels, and even busses/trams. Lots have limitations to hogging bandwidth though.

First thing that comes to mind is a mechanic’s stethoscope.

like this

Edit: basically 8adger’s screwdriver trick but I have one in my Kit of Resourcefulness™

That’s the route I took too. NAS for storage and simple docker containers, Minipc for compute/GPU.

Bitwarden/vaultwarden is a popular option for selfhosters.

That’s good news! It would be great if relays made it difficult to be targeted. I last tinkered with TOR almost… Jeez!.. 20 years ago haha!

I ran a relay too way, way back in the day and I remember almost a third of the sites I used blacklisted my IP address within days. It wasn’t cool.

I ended up shutting it down, resetting my cable modem, and spoofing a new MAC address on my router to get a new IP address to get everything working again.

Using a VPN is smarter. I wouldn’t run that on IPv6 whatsoever.

Every Plex client is a little different, but there is usually a video details or “playback info” button that will give you stream info such as direct play, transcode, or transcode (HW) for hardware support.

I just did something sort of like what you are doing and after a few hiccups, it’s working great. My Synology just couldn’t handle transcoding with docker containers running in the background.

Couple differences from your plan: I chose a N100 over the N150 because it used less power and I wasn’t loading up CPU dependent tasks on the thing. The N150 is about 30% faster if memory serves, but draws more power. Second, do you really need a second m.2 SSD BTRFS volume? Your Synology is perfectly capable of being the file storage. I’d personally spend the money you’d save buying a smaller N150 device on a tasty drive to expand the existing capacity then start a second pool from scratch.

Finally, I wouldn’t worry about converting media unless you are seriously pinched for space. Every time you do, you lose quality.

Ditto to your comment except power usage. I moved my Plex/Jellyfin (and hopefully Immich soon) docker containers to an N100 for the hardware acceleration. TDP is 6 watts on some of these devices and CPU use sits around 2% unless Plex is doing DB optimizations (about 60% for a bit). I haven’t measured consumption or my older server, but I feel moving some CPU intensive services to hardware GPU is saving a few watts.

There is M.2 on the mobo so I’d probably go with NVMe over SSD.

I second the RAM recommendation. I have 32GB in my Synology and it needed it for all those docker containers and VM’s.

As for the mobo, not thrilling, but could work. If you have to add a PCIe card for more drives, 10G network, or more NVMe, you’ll max out pretty quick with a GPU in there too.

Totally get it. And if you find a cool solution, let us know.

I’ve been slowly, very slowly, migrating away from Synology stuff, but everything you mentioned are my holdouts because they have been rock solid for decades. Even the cheap used products can swing those apps.

I hate to be the bad influence (no, who am I kidding, not really) and suggest more servers, but if you can find something cheap, I’d maybe give it a try.

I tried to update my lemmy instance and it all went so horribly wrong. DB never came up, errors everywhere, searching implied I updated to a dev branch sometime in the past (not a dev, don’t think I did) and it’ll be console and DB queries for a fix.

Ran out of time and overwhelmed, I restored backups and buried my head in the sand. Nope, not now. Future, yes, but oh not now.

Same here. I don’t like some of the recent decisions, but I remember the time I looked at the value and thought “yeah, this is working, valuable, and I can get behind it”, and bought the lifetime pass.

And I used the hell out of it! I don’t regret supporting the developers at all.

But features like plugins disappear, rolled to in-house teams. They work better, but cost more to maintain.

It’s ambitious, and gives developers plenty of work, but I feel the new redesign bit more than they can chew and overran budgets. They may be trying to balance budgets.