What’s up, what’s down and what are you not sure about?
Let us know what you set up lately, what kind of problems you currently think about or are running into, what new device you added to your homelab or what interesting service or article you found.
I finally finished my first iteration of my Minilab including a very smooth migration from the old server yesterday so I can go to the service side of things again. I plan to get some kind of selfhosters VPN for external access to stuff that’s not exposed to the internet, I’ll have to investigate which one.
Configured changedetection.io to notify me when my usual bus is delayed or canceled.
Latest thing is my server was hard locking up randomly every couple days. Finally thought to check IPMI and it was triggering a correctable ECC error on a specific stick of RAM.
I figured maybe the first couple errors were correctable by the ECC RAM but then they just got worse and caused the lock up.
Pulled the 2 sticks in that pair and so far so good. I’ll survive just fine with the remaining 192GB of RAM lol.
Also switched from my old Dell box with Opnsense to a Linksys MX4300 running OpenWRT, saves me about 20W and its fun to try something different.
Hello! I recently deployed GPUStack, a self-hosted GPU resource manager.
It helps you deploy AI models across clusters of GPUs, regardless of network or device. Got a Mac? It can toss a model on there and route it into an interface. Got a VM on a sever somewhere? Same. How about your home PC, with that beefy gaming GPU? No prob. GPUStack is great at scaling what you have on hand, without having to deploy a bunch of independent instances of ollama, llama.ccp, etc.
I use it to route pre-run LLMs into Open WebUI, another self-hosted interface for AI interactions, via the OpenAI API that both GPUStack and Open WebUI support!
Oh that’s dope. How many hours are you running? Do you also use them for things like encoding or something like that?
I’m working on self-hosting my own LLMs.
I realized there are things I wanna talk about and research but I don’t want to send it to open AI. Frankly I feel gross about how much I’ve sent to open AI. My desktop is a beefy gaming rig that I don’t use for gaming much. I have a 20thread core, 64gb ram, an Nvidia gtx 3060 and 5 spare TB so why not.
- I keep a few ollama models downloaded and I’m slowly getting to know them and what they can do. Gemma seems to answer the fastest so I’ve been using that. Deepseek is like the reasoning button on chatgpt.
- I use openai-whisper to transcribe meetings I record using OBS. It’s really slow so I have a cronjob transcribe all my meetings for that day overnight.
- Open Web UI is a fantastic LLM frontend. It provides tools, rags, web searching, and model ranking all as a simple to use UI.
- My desktop has a Wireguard server which makes it easy to use my OpenWebUI on my phone.
Now I want to work on giving the LLM access to my Google calendar so it can create reminders for me. I’m sick of forgetting to think about remembering to do things so I hope if I can just ramble at the LLM about what I’m doing or what’s on my mind it can organize my thoughts. What else are these LLM actually for?
Today I learned that for some reason some DNS servers don’t like SRV records, so had to troubleshoot it when people were unable to log onto my Minecraft server that is on a non-default port.
Having electric stability issues this week in Bangkok - several 2-3 hour outages, which are too long for a UPS to cover the gap. I have several mid range but older PCs running docker, virtualbox, etc for various things including a postfix server for the family email, immich, QBittorrent, pihole, paperless, huly, postiz, a Minecraft bedrock server, a flightradar24 ads-b collector, and a variety of other homegrown projects.
Thinking about getting some or most of this over to a service like hetzner, perhaps even splurging on a baremetal dedicated system.
Recently I’ve been reading about/trying to learn qemu and proxmox, but don’t understand them yet. Is that where it’s at for managing a bunch of your own VMs? Or kubernetes/k8s?
I’ve been a little out of the loop for a few years and of course coming back up to speed IT wise judge take weeks. Looking for recommendations on offloading my home stuff to a cloud that I control.
Proxmox runs Qemu under the hood. It’s the current favorite for VM management.
I wouldn’t bother with k8s unless you’re deploying services in high availability, or groups of related containers.
proxmox
You will enjoy Proxmox. When you get it all jammy, check out the Proxmox Helper Scripts: https://community-scripts.github.io/ProxmoxVE/
Hey that’s awesome! thank you for the share. Planning to install proxmox this weekend and give it a try.
Thinking about getting some or most of this over to a service like hetzner, perhaps even splurging on a baremetal dedicated system.
If I may, I find LUXVPS to be quite capable and responsive hosts.
Black Luxury Deal #1
4 vCores (Xeon Gold 6150) 26 GB DDR4 RAM 150 GB Raid 1 NVMe 1 Gbit internet speed | 40 TB Traffic 1x IPv4 1x /64 IPv6 3.2Tbit Premium DDoS Protection 24/7 Ticket Support 4 Backups For ONLY 10€/Mo (recurring)I’ve never used Hetzner, and I don’t know what you are hosting, but I’m sold on LuxVPS. I also use Contabo, and Ethernet Services. The latter would indeed be bare-bare-metal as there are no frills. However, for a test server and for $35 a year, it works.
I’ve used a RV/Marine deep cycle battery attached to a UPS before, that would certainly give you enough for 2-3 hours on most setups.
K8S is a whole different approach and I find it to be a lot more complex, but you would not need virtual machines. If all your applications are running in containers anyways, you could consider it. Finding a good solution for persistent storage is probably the most important design decision.
Not really self-hosted, but I set up obsidian with syncthing and am going to transfer all of my notes from book stack to it and let bookstack be more organized documentation and obsidian to be a big scattering of notes and tags and such. I tried it with bookstack, but the flow was too much of a barrier for me to use it consistantly
For some reason Grafana started to sync roles with my IdP (google) and now my own user keeps getting a read only role, so I decided to take this opportunity to finally move away from google and start hosting keycloak instead.
It was a busy week so I could not get the time to finish it yet.
You may also have a migration path by hosting keycloak and add Google as an Identity Provider. Gives you much more flexibility and control this way
Thanks for the tip! I didn’t know that setup was possible.
Sweet!
What’s up is everything I’ve been running and down is what I haven’t.
not working
I haven’t been able to get friendica to connect to Maria DB, so I’ll eventually try just MySql. Grafana isn’t running bc I would need to change a lot of things to get an exporter into each container and the truenas apps don’t really allow that configuration - fine if you have docker compose though, which I’ve started doing more and more.
new
I just got up and running with Stirling pdf, a free (and paid) PDF editor. That looks pretty sweet.
But I’m now also using 15GB of the 32 on the system, which is still plenty for Arc cache for me
what I want
I want to rent a VPS to host various fediverse apps, probably Lemmy, pixelfed, and write freely to start, for the nomad/expect communities. I’ve been looking at netcup and they have some decent arm offerings.
I’d like to put Talos Linux on it so I can get some kubernetes experience. They have a good sized server for €10, so I could expand to add a DB server or one specifically for logging and metrics.
I was looking at Hetzner, but I’ve read that their block storage is super slow and causes timeouts on DB.
Of course, can I even run these apps on arm? I guess I gotta find that out.
One thing I’d like to do is make a web page that makes signups super easy and would create an account on all services, ideally. Not a huge deal of that isn’t reasonable, but it’d be nice to allow doing it once rather than multiple times. If I could get sso, that’d be good, but I don’t know how supported that is.
I’ve installed coraza web app firewall with OWASP ruleset this weekend. I must admit that it wasn’t as easy as I expected it, but it now (mostly) works. I had to give up with nextcloud though.
I’m currently trying to figure out why my email server got blocked by Proofpoint and they refuse to talk to me. Really about ready to give up on email after self-hosting it for a decade with few problems.
Oh that sucks! One would think that after that long, it’d be somewhat established.
RIGHT?!
Check RBLs a lot of times services just use one of those, and they can be flaky. Usually, you can fill out a form and get reinstated.
I’m not on any of those blacklists, luckily. I guess Proofpoint doesn’t publish theirs. At least iCloud and Gmail both use them. I saw one hint that they may require mail servers to literally have the word “mail” as the subdomain, so I’m working up the courage to mess around with my perfectly working DNS.
There is still the relay through the cloud route (SES, but also at least Scaleway)
Part of me thinks if I have to pay for a relay service, I should just pay for hosted email. But I’ve definitely been considering it!
I need to get a new VPN setup. Been using OpenVPN through OPNsense for years but I’m fed up with the abysmal performance of the OpenVPN client on iOS. Open to suggestions but it has to be fully self hosted.
Wireguard is where it’s at.
Good on iOS too, albeit a bit battery hungry if you route 0.0.0.0/0 and ::/0
I dunno if there’s an iOS equivalent but on my Android phone I use the WG Auto Connect app so it’s only active when not on my home wifi.
The iOS app has this, based on SSIDs
wg-easy is what you want
I have running OPNsense as well and was looking for OPEN VPN on it as well, but I’m not decided yet. I wonder if Android clients are any better.
I’m running OpenVPN on pfsense and am using the android app.
I’ve got a stable 150mbit/s, depending on carrier coverage.
DOWN:
I’m currently fighting with my OliveTin config file. I added a simple new config for a button action and ylthe whole thing just shit the bed. Now OliveTin won’t load at all. Even after removing the new config. Stupid yaml.
UP:
After reading the Jellyfin docs and their Hardware Encoder Quality section which states
Apple ≥ Intel ≥ Nvidia >>> AMD*
I decided to spin up a test server on the m1 mini that’s been sitting unused in my basement for a couple of months now to see if I can get better performance out of jellyfin on the m1 vs where it’s running currently, which is on an i7 Intel that’s going on 10ish years old now.
I also spun up baserow and directus containers to see which one I want to use for my database needs.
I started this about a month ago, absolutely no idea what I was doing, and in that short time this little box has grown a ton. Got the basics for cloud storage, jellyfin with the arr suite, navidrome to replace spotify/tidal, etc. Got my scanner going right into paperless, finally starting a budget planner with actualbudget, even set up homebox to maybe eventually keep track of my collections of random bullshit. Spent 3 days fighting with Wireguard and gluetun to make a single VPN connection that’ll hook me into my LAN but also output all my traffic through Mullvad, using pihole as my DNS - I should get Unbound set up at some point too but that’s a project for another day.
Today I learned about homeassistant, and while I’m not one to care about IoT shit or whatever, just dabbling with NFC tags for the lights and such has been pretty neat.
This week I’m getting a second machine in that I’m going to use exclusively as a NAS and stop relying on USB external hard drives.
I really just wanted a little 24/7 Bob Ross box with a bit of cloud storage, and this project blew up a lot more than I thought it would LOL
Bob Ross box???
Grab the entire series, load it up on the tv, and let it rip all day
Anyone have a good guide on setting up a reverse proxy that works with tailscale? Not sure if there’s anything specific I need to keep in mind or if it would just be setting up the reverse proxy like normal. Thinking of using either traefik or caddy.
It should be the same setup regardless if you’re using a VPN or not.
Having used both I generally prefer traefik.
You’re gonna need to provide more detail on what you’re trying to do
Might look into the pangolin project if what you’re trying to do is expose services from your home network over wireguard to a reverse proxy on a vps.
The software suite is basically wireguard, traefik, and auth middleware wrapped in a trenchcoat. Much simpler than rolling your own implementation, but there has been recent controversy with the project over locking “basic” existing features behind a paywall after the project got popular, though after public backlash they’ve backpedaled on that iirc.Edit: Just realized you said tailscale. Above recommendation might be a deal breaker depending on your reason for wanting tailscale specifically
All good, thanks for the recommendation. I’m using tailscale as I currently don’t want to expose anything over the Internet and also don’t mind tailscale being a freemium service. I might still look at pangolin just to expand my knowledge.
You can restrict Caddy access to use your tailscale. For instance in your Caddyfile:
For tailscale ip range:
myverycoolserver.duckdns.org { @allowed { remote_ip 100.64.0.0/10 # Allow Tailscale IP range } respond @allowed 200 # Allow access respond 403 # Deny access for others reverse_proxy localhost:YOUR_SERVICE_PORT # Your service configuration }For specific tailscale IP:
myverycoolserver.duckdns.org { @allowed { remote_ip YOUR_TAILSCALE_IP # Replace with the specific Tailscale IP } respond @allowed 200 # Allow access respond 403 # Deny access for others reverse_proxy localhost:YOUR_SERVICE_PORT # Your service configuration }I have caddy on a vps that serves as a tailscale exit node and also reverse proxies over the tailnet. My pfsense router is also in the tailnet and exposes some subnet ip addresses to the tailnet. So for example I have public domain watch.example.com hits my caddy and gets proxied to internal IP 192.168.31.48 which is my jellyfin docker.
