from tofu@lemmy.nocturnal.garden to selfhosted@lemmy.world on 16 Nov 09:46
https://lemmy.nocturnal.garden/post/360803
How’s your stuff doing? Unplanned interruptions or achieving uptime records?
I’m currently sailing rather smooth. Most of my stuff is migrated to Komodo, there will stay some exceptions and I only have to migrate Lemmy itself I think. Of course that’s when I found a potential replacement but I’ll let it sit for a while before touching it again. Enjoying the occasional Merge Request notification from the Renovate Bot and knowing my stuff is mostly up to date.
I’m thinking about setting up some kind of Wiki for my other niche hobby (Netrunner LCG) lore as there’s a fandom one that most people avoid touching and updating but since I likely won’t have time to start writing some articles on my own as a kickoff I’m hesitant. Also not sure which wiki I’d choose as well.
threaded - newest
Currently working on moving the more family-relevant services to OIDC-based login via Pocket ID passkeys so I can put my parents on them.
Also, still on the lookout for a good Nextcloud replacement. Even Opencloud displays the first signs of feature creep.
What’s wrong with Nextcloud?
It grew from a nice Owncloud fork into a do-it-all groupware solution by adding on more and more things without really improving the basis. Each version the performance gets a little worse, syncing gets stuck more often, etc.
Opencloud looks or at least looked good as it started out as an Owncloud Infinite Scale fork, but of course they’re adding on more and more groupware stuff without improving the core first. Maybe we’re doomed to witness the same cycle with each solution, who knows.
Also, Nextcloud/Owncloud are written in PHP, that might also have a significant impact on its poor reliability ^^
When I ran Nextcloud, it broke every other update. Mostly because NC didn’t seem to care that anyone had a 7-year-old install being migrated along.
Aside from being hella slow, I just don’t like that it can’t use the same directories as my network shares and requires uploading. This script might help but honestly I just stick to the basic shares because of this
Purchased 5 1tb drives to expand my study server. Going from 600GB to 4TB is going to make more complex labs possible.
Certainly not my homelab as my server isn’t booting since a few weeks ago and I didn’t fix it yet…
Oh no!
I recently installed Beszel and really like it but I would prefer not to have to login every time I want to check my systems. Is there any easy alternative?
Idk Beszel, but generally you could check if you can increase the session expiration time in the config or put it behind some SSO like authentik
I use NetData, with the v3 ‘switch’ on the url. Example: netdata.mycoolserver.com/v3. The v3 lets you skip the login process and head right to monitoring observables. Some people may have concerns about NetData, however it covers just about every metric I think one would need, all in one package.
Chose yesterday late evening as the time to migrate my containers from docker to podman (still rootful). By luck most things work again, except wireguard/qbittorrent
What made you want to switch from docker to podman?
Nice. I’m aiming to go from bare metal to rootless podman managed via quadlets. Networking seems like the difficult part.
Planning to host a Nix caching server, and have CI build all package and NixOS outputs on every push to git, then in turn pushing the output artifacts to the cache. Would save me a good chunk of time when tinkering with VMs that haven’t seen manual updates in a while.
Only thing is, I’m not sure how to approach building and caching NixOS configs that receive agenix secrets in their input. Obviously those should not be cached…
I finally moved my mail server from Hetzner to my homelab.
Pretty smooth sailing so far. For now I’m using Scaleway for outgoing mails since I can’t set a PTR record here but I might just try sending a few without PTR to see how other providers react.
From my experience using a mailserver with no PTR and an ISP who likes to put their addresses on a PBL, it’s very good. Gmail tends to be the most annoying and wants that PBL listing removed or you’ll go to spam for new recipients, but other than that 10/10. I’d be interested to hear what your findings are if you do test it!
@domi @tofu A thought but is there a way to exchange trust with other mailservers to guarantee delivery?
Just got some power measuring plugs. Home Assistant and immich-running raspberry pi + NAS (dual 20TB in raid 1) + switch clock in at around 30W. Surround receiver playing music ups that by 90W. After a minor water leak I added 5 leak sensors to the system that will blink lights and send texts if they detect anything.
The biggest problem is that I’m still running lights through hue and some of them have an annoying tendency to drop off the network…
Get yourself a Sonoff ZigBee bridge! Hue light support is practically native, and they act as extenders to reach your other ZigBee devices! Just don’t expect to be able to sync them with any movies or peripherals. I think there is a virtual Hue bridge on HACS and that might help with that, but idk
I have that. I just got hue first, so all my lamps (or at least the old ones) are registered in hue. I haven’t taken the time to move all of it over, so now I have two competing networks.
I’ve finally setup Netbird instead of Tailscale to VPN to my network. Took some time since I wanted it to work with pocket-id and had some issues configuring everything properly. Runs like a charm now.
I’ve just finished to configure my homelab with wg-easy yesterday to do exactly that. Took me weeks because podman. And now I learn that there was a better way? Oh well…
I updated my Dietpi setup today, because a new version was available. It went very well, and everything works perfectly after a reboot.
I always hold my breath whenever I’ve done anything major to the server and I need to reboot.
Right? It’s like a trust fall. You just have to cross your fingers and hope for the best.
I recently switched my phone from Android to GrapheneOS and now rely even more on my selfhosted services. Immich is such a great project. Still gotta figure out my music collection though, since switching from YT Music to Jellyfin. Most of it is sorted by date of purchase, because that worked best with my DJ workflow. Now I gotta bring it over to a folder structure that works for jellyfin. It seems like the answer is musicbrainz Picard, but I gotta figure out how to configure it.
Also been thinking about some AI ideas I’d like to try, but I have zero intention getting involved with openai, meta, google or whoever the fuck. So self hosting it is. But on what hardware? Option 1 seems to be to get some professional server board, CPU, ram and start with one RTX3090 and go from there with the option to hook up more GPUs. But a setup like that sounds like it would cost some serious money in electricity. Option 2 seems to be a Rzyen AI Max+ 395, configured with a fuckton of ram, available to the whole apu and as suchs usable for memory hungry models. This seems to be much much more power efficient. But its all integrated and I couldn’t swap out components or upgrade in the future. Leaning towara option 2 atm, but maybe I’ll just wait a bit longer and see what else comes up in the coming months.
Nice… I use ytdl-sub for downloading music, highly recommend it. You can write tag metadata but if you want embedded stuff I’d recommend trying beets. Running both as a user whose primary group matches Jellyfin is a must if you want stuff saved next to the video files… The dev is also very active.
I just installed Ollama and use gemma3 for now. I wanted to use dolphin-mixtral but holy crap it wants more RAM than my entire setup
This is basically what I’ve found with self hosted AI. I just don’t have the equipment for it. Would love to be able to host a selfcontained LLM, but alas, as you say, it eats up resources. FEED ME MAURICE!
I ended up using navidrome for my music to take advantage of the subsonic API, which has been phenomenal 🙂
I’ve just finished ripping about 1300 CDs. I used Jellyfin for a bit since I already had it set up for my video library, but I wasn’t happy with the Android options and it was pretty basic.
Navidrome is a fucking TREAT. Paired with Symfonium, I’m finally enjoying my personal music collection regularly again.
As for tagging, OP, while I get why people like Picard, it doesn’t always work with how I like to do things. I put everything into a music folder on my desktop, use Mp3Tag to retrieve metadata, edit what I need, and make sure the artwork is decent and sized where I want it. Then I use the tag > filename to organize and move them to my NAS.
I’ve found Navidrome to be quite capable of handling large music collections. I was worried in the beginning. It sips resources. When I fire it up and listen remotely, I watch the CPU and RAM. It barely moves the needle. Very happy with it.
Trying to host stable diffusion to generate some art for my D&D campaign.
There are a lot of opensource virtual tabletop gaming platforms that really look nice. I used to be heavy into D&D back in the day.
Anything as good as Foundry?
Well, Foundry is a standard, can’t deny that. LOL You might want to compare that to Caldron VTT
I recommend ComfyUI. It makes running everything trivial, and is very easy to learn, use, and extend.
I also recommend supporting artists directly and learning to draw.
Thanks. I support artists when I can for art I intend to share at the table. The AI is just for me to easily reference characters.
Smooth sailing for me too, shockingly. I’ve recently added my 26th service to Proxmox - LibreELEC (Kodi), with the very complex matter of monitor passthrough. It’s such a versatile program and it has replaced my Chromecast with more features and side bonuses than I could’ve imagined. Another huge step towards degoogling.
I wish someone would jailbreak the Google home and Chromecast devicea so we don’t have to throw them away in a year when Google abandons them.
Right? My one has developed a nasty case of defective WiFi/Bluetooth chip. I’m convinced it’s a superficial, intentional break. Flashing it could well revert what Google’s doing to it
I have been experimenting with a btrfs raid array and am getting some new hard drives in the mail today, hoping it goes smoothly and they work 😬 All part of a larger goal of migrating my synology NAS to a purpose built machine.
Also got my first contribution and donation on my OIDC SSO project, which is really exciting!
Ey! congrats for the donation. I hope your personal project succeeds!
Just installed Owncast, so townsfolk can ride my G-scale Polar Express via an onboard livestream, as part of a revamped lighting and projection mapping festive season show.
While I was at it I also added Kokoro for TTS.
Thought I would spice up Jellyfin for the festive season, so am trying out the Jellyfin Enhanced and Home Sections plugins.
Going to try to convert two 2-post racks into a 4-post rack today. Dreading the mess though.
Switch my ActualBudget authentication to my locally hosted Authelia instance and it was a breeze, with no issues. Always a fantastic feeling.
Working on automating tasks so I don’t have to block out hours of time a week managing everything. Just got watchtower running and going to see how it does before trying out some other automations.
If you find that watchtower (original) screws up the updates frequently there is a watchtower fork that runs so much smoother. I don’t have any issues with it at all. The original watchtower app hasn’t had an update in 2 years, so it might be something to keep in mind.
In fact you must use the fork. The old one no longer works with recent Docker, due to API versioning. I found that out last night when I brought up my compose stack and traefik wouldn’t start, because it too needed an update.
I had that issue with Portainer recently. I had to drop back to the previous docker version, and held it until Portainer works through the snag. I didn’t think about original watchtower being affected. I just got tired of having to fix broken updates, and went looking for something better. When original watchtower worked tho, it worked well.
I’m actually using this one which seems to be more actively maintained than the one you linked.
Bookmarked! Thanks for that. Learning all kinds of stuff today.
A recent t480 purchase may replace my second workstation tower, which I think is about to become my most powerful server in the cluster….
So nothing new hosting-wise, but that tower I can shove the spare 12tb and 4tb drives I have and net myself another 30ish TB’s of usable storage, more once I replace the 12TBs in one of my NAS boxes with 18tb or more.
Speaking of which - where the hell do I track prices these days? diskprices.com seems to be a mess of inaccurate pricing and shucks.top can no longer track even half of what they used to. What a mess.
PCPartPicker is your best bet (hint: sort by price/gb), but they don’t really track shucking prices
Honestly with what Seagate has been doing with their externals, shucking is probably best avoided at this point.
That said, yeah, seems like and its not a perfect option either, seems like I’ll have to use multiple sources and just keep an eye out with a daily check or something.
Everything here is smooth sailing. I have been trying to track down a bothersome Suricata entry.
ad nauseum. There are three individual ips. One from Singapore, one from China and one from Romania. They are being blocked, so that’s good. Thing is, these are from realitvly ‘clean’ sources:
120.132.37.195 was not found in our database202.136.163.11 was found in our database! This IP was reported 5 times. Confidence of Abuse is 0%:On the server side, I have nothing calling out to these ip. That’s what was really bugging me. Nothing server side, just these three bothersome ip hammering Suricata. Generally, I would dismiss as benign and part of normal UDP behavior. However, it’s the constant hammering that makes me suspicious. Could be high volume port scanning. However, it could also be known attack campaigns like UDP amplification attempts.
Other than that, I might find something to get into today.
Switching my main PC to nixos from fedora atomic sway. The sway config tripped me up last time, this time I’ll succeed! (I hope)
I started out rewriting my network backup scripts only to realize I was adding functionality to a previous script I wrote to automatically mount and dismount luks encrypted volumes. I still want to type in my luks passphrase because I don’t want everything automated and prefer to include inconvenience as an additonal security measure in securing some of my data.
I also came to the realization recently that the reason I don’t relate strongly to other self hosters is because I’ve unknowingly been trying to create a minimal self hosted system that is more beneficial to small, low powered devices.
I’ve been using Alpine Linux, I install only the bare, older but well established tools and have been creating scripts soley based off those tools instead of seeking out bigger, more complicated modern tools. For example creating workflows by only using
rsyncor using github.com/RayCC51/BashWrite to create a blog that only usesbashand GNUsedto create a static blog site.At least now that I’m aware of this, I can keep an eye out for such projects or communities and would hopefully be able to contribute something in that direction.
There’s absolutely nothing wrong with minimal. The way technology is in this timeline, you really don’t need a lot to get a lot out of it.
Your perspective aligns with a lot of self-hoisters who run things on rpi’s and such, but not the “home labbers”. Also, see the pubnix, tildeverse, smol web, indie web, and to some extent the retro computing communities. You are definitely not alone!
I actually started with RPi’s. The first one, a used Pi 4b, is dedicated only to HomeAssistant. I don’t tinker with it anymore because it does what I want and I don’t want unexpected downtime when I have to use the bathroom or use the lights in my room.
I bought a used Pi5 with the intention of upgrading later. In life I am quite minimal and find a joy in using what little tools and material I have to create something new. That seems to hold true to technology and scripting too. The RPi5 with an old USB3 HDD is actually way more power than I can currently use and can imagine using for a long time. The extra room to work is convenient though.
I’ll have a look into some of the places you suggested, those seem like the places to draw good inspiration from, thank you.
Trying to work up the courage to troubleshoot a very worrying disk error on the new NAS I’ve been building, which if solved will leave me the problem of working up the courage to try and migrate to the new server without losing my Plex library settings and progress.
Basically I’m frozen in fear.
Trying to run a fediverse server on a decade-old Wi-Fi router and encountering some
unexpected issues. Making progress, though.Sounds cool, which software are you using?
I installed Jellyfin on my server and threw kodi on a minipc I dug out of dumpster pile at work. Works pretty well, but my server needs more RAM and the minipc needs either a wireless keyboard or a USB-HID remote controller to finalize the setup. Also ran some wiring in the house and added two network sockets to a room where the whole kodi-tv-gamingpc-whatever-pile is going to live.
On the server RAM I found some on ebay, but if anyone is interested on 64G DDR4 ECC DIMMs I have a few. I thought they were supported on my server motherboard when I took them out from a old server at work but it supports only up to 32G ECC dimms.
Hi! I installed LibreELEC in a RPi4, and connected to the hdmi of my TV I can control it with the atV remote, I don’t know if it will usefull…
I’d rather have a physical remote which acts as a keyboard so it’ll support waking the system up from suspend. Plus I prefer a dedicated device for that instead of a phone as I’m not a only user for the thing. There’s plenty of those around, only problem is to find one that works reliably and local stores don’t seem to have a lot of options so I might need to dig one up on ebay even if it’s a bit of a PITA to order from China to EU today with customs.
I’m pretty sure they’re referring to hdmi-cec, nothing to do with a phone.
I didn’t know raspberry supports that. Searching for ‘atv remote’ just brings up androind apps, so maybe I misunderstood. Neat thing, but the hardware I have doesn’t support it and seems like usb-cec adapters are more expensive than usb-hid remotes.
Tried to setup a personal matrix server last night, got it to federate, next step is Matrix’s Element Call, spent too many hours trying to block the
/_synapseendpoint with Traefik because it is recommended by Matrix, no luck unfortunately.All this in hopes I can add a Music Bot to my instance or something similar.
I just bought aan IP KVM switch for a hundo, now heading to the store foir a case of frosty’s and re-rack my servers to make room.
Finally got around to dns-01 and acme containers today. Hooray LE signed wildcard lab cert.
Mostly everything is running smoothly. Been fighting with some zigbee integrations randomly dropping connection from Home assistant but it’s nothing too important.
Biggest issue I’ve been facing is how to make sure all my media is properly encoded so jellyfin doesn’t pin my cpu transcoding when I’m streaming to the onn boxes around my house. Debating if I need to dump the onn’s and try to spin up raspberries for each TV instead
my server has been down for one week because I’m migrating to OpenBSD but I got a weird error while installing, but yeah, everything’s fine!
Had a productive session this weekend migrating my promtail config to grafana Alloy and setting up a syslog receiver to capture output from my cron jobs. Next up I’ll be messing with some scripts to sync my dashboard config across several instances which should be pretty neat if it works
I dug out an old laptop and installed Yunohost on it. I was so excited until I discovered that my ISP uses CGNAT. I’m trying to figure out what I want to do next.
I am looking at using headscale or just paying the US$10/month for a static public IP from my ISP. If I go with headscale, then it appears that I wouldn’t need Yunohost.
I’m a newb at this so there’s a lot I don’t know yet.
You can rent a cheap VServer as well and use its static IP to forward traffic. Easiest for it would be SSH reverse tunnel. Or you could VPN it with your homelab (connection established from within your homelab).
If you don’t want to rely on an external service you could as well establish a VPN server within your homelab and use IPv6 to connect to it, although the disadvantage would be, that if you’re trying to connect from IPv4 networks ‘outside’ that wouldn’t work.
Just listing some options to research. Welcome to the hobby, have fun 🤗
I’d rather not rely on an external service if possible. I’m just starting to read up on doing the whole VPN thing.
I appreciate your response and will keep your suggestions in mind as I move forward.
My ISP uses CGNAT but I can ask for a dynamic IP address for free. I sent them an email and got a reply in less than a week. I can also pay extra like 2.50€ per month or something for a fixed IP. I found that quite reasonable.
I’m thinking getting a static public IP might just be the easiest way to go. I have a pretty good ISP. Aside from sticking all the customers behind CGNAT.
Namecheap, and I guess other registrars too, has an API that you can call from your server to update your IP address in their DNS. It’s super easy. No need to pay for a static IP address. At least in my case ei already use my domain for other things.
And since when is the easiest way the funnest way? :P
I don’t understand how that’d work but I’ll look into it. Thanks for the info!
Basically it’s a URL that you call with curl. You can set up a crown job to call every day or as often as you need. The URL contains the domain name or subdomain, you dynamic public IP (not CNAT), and the API token. This way you Domain always points to your dynamic IP.
I think I get it. I’ll look into after I get home from work today.
Pretty smooth sailing at the moment. I’ve got:
All running on a 4 node raspberry pi kubernetes cluster.
One of my drives crippled itself a few days back, not sure what caused it. Wasn’t able to be resolved without a host restart which was unfortunate. SMART isn’t failing and has been working fine, so I’m chalking it down to a weird Proxmox bug or something.
For sure expected I was going to need to do a rollback on an entire drive after that restart though. Still may have to if it reoccurs.
Evening is going ok, but noticed the screen saver on jellyfin isn’t showing up lately… need to investigate…
Also, watched the latest “Explaining Computers” episode today.
docker-ce v29 update somehow messed up my homelab so badly that I had to downgrade to v28 to restore my system.
That was awful. Had to restore from backup.
Good to know 🫣
Following the FUTO guide, but having problems with getting mailcow going… I’ll hopefully figure it out by tomorrow.
Burning the midnight oil on my self hosted journal app: github.com/journiv/journiv-app
Oh hey I just thought about setting up a journal! Maybe I’ll check it out
Definitely!
Changed my family dashboard from magic mirror to a home assistant dashboard. I'm missing some cute things, but the major functions work better, and I get some options that I didn't before.
Made some changes to my I2P router today, but otherwise all good.
Bad week for me. Tandoor had become the home of quite a lot of recipes, and well, I’m never gonna just pull a docker container again without a backup, cause I did a pull and the bastard stopped working.
So I setup Django and got started doing my own recipe server cause I was never very enthused about Tandoor, too much netflix-like Presentation bullshit and did not allow for the very simple thing I wanted, which was, a compact list of my recipes by alphabet that I can swiftly click on the one I want.
I also need to get my Python chops back cause I think there will be jobs again, soon enough.
Meanwhile, anyone got any suggestions of a better recipe app? Needs to run as a Linux server, that’s about it. I can go Tailscale if it has no security. If I get mine to something usable I’ll make it available.
Trying to smoothly orchestrate prowlarr, radarr, jellyfin, and transmission (via Proton vpn), using a big beautiful docker compose file. It’s been working OK but not without roadbumbs and tough learnings. Keep messing up directory permissions one way or another.
Next step is setting up fail2ban on my public facing jellyfin to control things a little better. Everything is hosted at home, and I don’t want to use cloud flare tunnels, are streaming video is technically not allowed in them.
If you have more good tips on securing a home server, let me know!
Also, this is all running on an ancient 2012 mac mini running Ubuntu. Slow as molasses and sometimes the fans make a noise. I should start looking into back-up solutions, at least for the configs.
Bought my first raspberry pi 5, 8 gb ram version. Gonna be using it to run a jellyfin server and maybe a foundry server if it can handle both concurrently. Anyone familiar enough to know if running multiple things on one of these is wise?