[acidburnNSA]

I have something like this, in the same case. I have beefier specs b/c I use it as a daily workstation in addition to running all my stuff.

* nginx with letsencrypt wildcard so I have lots of subdomains

* No tailscale, just pure wireguard between a few family houses and for remote access

* Jellyfin for movies and TV, serving to my Samsung TV via the Tizen jellyfin app

* Mopidy holding my music collection, serving to my home stereo and numerous other speakers around the house via snapcast (raspberry pi 3 as the client)

* Just using ubuntu as the os with ZFS mirroring for NAS, serving over samba and NFS

* Home assistant for home automation, with Zigbee and Z-wave dongles

* Frigate as my NVR, recording from my security cams, doing local object detection, and sending out alerts via Home Assistant

* Forgejo for my personal repository host

* tar1090 hooked to a SDR for local airplane tracking (antenna in attic)

This all pairs nicely with my two openwrt routers, one being the main one and a dumb AP, connected via hardwire trunk line with a bunch of VLANs.

Other things in the house include an iotawatt whole-house energy monitor, a bunch of ESPs running holiday light strips, indoor and outdoor homebrew weather stations with laser particulate sensors and CO2 monitors (alongside the usual sensors), a water-main cutoff (zwave), smart bulbs, door sensors, motion sensors, sirens/doorbells, and a thing that listens for my fire alarm and sends alerts. Oh and I just flashed the pura scent diffuser my wife bought and lobotomized it so it can't talk to the cloud anymore, but I can still automate it.

I love it and have tons of fun fiddling with things.

[VladVladikoff]

For anyone considering this, it's not a good plan to do it this way, if you have any family members relying on these services, you have to kill them all every time you reboot your workstation. It's really not great to mix destop and server like this. (speaking from experiance and I really need to get a separate box setup for this self hosted stuff)

[zem]

> if you have any family members relying on these services, you have to kill them all every time you reboot your workstation

yikes!

[Pooge]

Yeah I can't imagine killing my family members every time I'm shutting down my computer

[altano]

It's better than having to hear them complain every time plex goes down

[giwook]

And yet sometimes you just need to pull the plug.

[bjackman]

You are always gonna have some downtime in a homelab setup I think. Unless you go all in with k8s I think the best you can do is "system reboots at 4AM, hopefully all the users are asleep".

(Probably a lot of the services I run don't even really support HA properly in a k8s system with replicas. E.g. taking global exclusive DB locks for the lifetime of their process)

[embedding-shape]

> You are always gonna have some downtime in a homelab setup I think. Unless you go all in with k8s I think the best you can do is "system reboots at 4AM, hopefully all the users are asleep".

Huh, why? I have a homelab, I don't have any downtime except when I need to restart services after changing something, or upgrading stuff, but that happens what, once every month in total, maybe once every 6 months or so per service?

I use systemd units + NixOS for 99% of the stuff, not sure why you'd need Kubernetes at all here, only serves to complicate, not make things simple, especially in order to avoid downtime, two very orthogonal things.

[bjackman]

> I don't have any downtime except when I need to restart services

So... you have downtime then.

(Also, you should be rebooting regularly to get kernel security fixes).

> not sure why you'd need Kubernetes at all here

To get HA, which is what we are talking about.

> only serves to complicate

Yes, high-availability systems are complex. This is why I am saying it's not really feasible for a homelabber, unless we are k8s enthusiasts I think the right approach is to tolerate downtime.

[embedding-shape]

> So... you have downtime then.

5 seconds of downtime as you change from port N to port N+1 is hardly "downtime" in the traditional sense.

> To get HA, which is what we are talking about.

Again, not related to Kubernetes at all, you can do it easier with shellscripts, and HA !== orchestration layer.

[furst-blumier]

I run my stuff in a local k8s cluster and you are correct, most stuff runs as replica 1. DBs actually don't because CNPG and mariadb operator make HA setups very easy. That being said, the downtime is still lower than on a traditional server

[ryukoposting]

It's also worth noting you don't need sophisticated hardware to run anything listed in the parent comment. 8GB of RAM and a Celeron would be adequate. More RAM might be nice if you use the NAS a lot.

[wbjacks]

Have you tried using snapcast to broadcast sound from your Samsung tv? I gave it a shot and could never get past the latency causing unacceptable A/V delay, did you have any luck?

[pajamasam]

Impressive that all that can run on one machine. Mind sharing the specs?

[c-hendricks]

I run similar (gitea, scrypted+ffmpeg instead of frigate, plex instead of jellyfin) plus some Minecraft servers, *arr stack, notes, dns, and my VM for development.

It's an i7-4790k from 12 years ago, it barely breaks a sweat most hours of the day.

It's not really that impressive, or (not to be a jerk) you've overestimated how expensive these services are to run.

[hypercube33]

Video is usually offloaded too to the igpu on these. I have like 13 vms running on a AMD 3400g with 32gb

[pajamasam]

Fair enough. How much RAM though?

[decryption]

16GB would be plenty. I've got like a dozen services running on an 8GB i7-4970 and it's only using 5GB of RAM right now.

[shiroiuma]

If you're running ZFS, it's advisable to use more RAM. ZFS is a RAM hog. I'm using 32GB on my home server.

[renehsz]

ZFS doesn't really need huge amounts of RAM. Most of the memory usage people see is the Adaptive Replacement Cache (ARC), which will happily use as much memory as you throw at it, but will also shrink very quickly under memory pressure. ZFS really works fine with very little RAM (even less than the recommended 2GB), just with a smaller cache and thus lower performance. The only exception is if you enable deduplication, which will try to keep the entire Deduplication Table (DDT) in memory. But for most workloads, it doesn't make sense to enable that feature anyways.

[hombre_fatal]

That + full-disk encryption is why I went with BTRFS inside LUKS for my NAS.

They recommend 1GB RAM per 1TB storage for ZFS. Maybe they mean redundant storage, so even 2x16TB should use 16GB RAM? But it's painful enough building a NAS server when HDD prices have gone up so much lately.

The total price tag already feels like you're about to build another gaming PC rather than just a place to back up your machines and serve some videos. -_-

That said, you sure need to be educated on BTRFS to use it in fail scenarios like degraded mode. If ZFS has a better UX around that, maybe it's a better choice for most people.

[c-hendricks]

32gb for me because half of that is given to the development VM

[TacticalCoder]

> Impressive that all that can run on one machine. Mind sharing the specs?

Not GP but I have lots of fun running VMs and lots of containers on an old HP Z440 workstation from 2014 or so. This thing has 64 GB of ECC RAM and costs next to nothing (a bit more now with RAM that went up). Thing is: it doesn't need to be on 24/7. I only power it up when I first need it during the day. 14 cores Xeon for lots of fun.

Only thing I haven't moved to it yet is Plex, which still runs on a very old HP Elitedesk NUC. Dunno if Plex (and/or Jellyfin) would work fine on an old Xeon: but I'll be trying soon.

Before that I had my VMs and containers on a core i7-6700K from 2015 IIRC. But at some point I just wanted ECC RAM so I bought a used Xeon workstation.

As someone commented: most services simply do not need that beefy of a machine. Especially not when you're strangled by a 1 Gbit/s Internet connection to the outside world anyway.

For compilation and overall raw power, my daily workstation is a more powerful machine. But for a homelab: old hardware is totally fine (especially if it's not on 24/7 and I really don't need access to my stuff when I sleep).

[leptons]

Cheap to buy old hardware, but electricity to run those old rigs isn't really cheap in many areas now. My server is costing me about $100/month in electricity costs.

It does have 16 spinning disks in it, so I accept that I pay for the energy to keep them spinning 24/7, but I like the redundancy of RAID10, and I have two 8-disk arrays in the machine. And a Ryzen-7 5700G, 10gbit NIC, 16 port RAID card, and 96GB of RAM.

[shellwizard]

It depends on the type of hardware that you use for your server. If it's really server grade you're totally right. For example cheap memory+CPU+MB x99 off AliExpress are cheap but they're not very efficient.

In my case I fell in love with the tiny/mini/micros and have a refurbish Lenovo m710q running 24/7 and only using 5W when idling. I know it doesn't support ECC memory or more than 8 threads, but for my use case is more than enough

[gessha]

I’ve been watching some storage and homelab-themed videos and I heard there’s a lot of optimizations you can do to lower power usage - spinning the disks down, turning the machine on for a limited time, etc.

[leptons]

That doesn't work for me. The main server is constantly using the disks to record security cameras, run VMs 24/7, Plex, a web server, a VPN (so I can dial in to my local network remotely), and a lot more.

[matja]

How have you measured the power usage/cost? That seems like a incredibly high price for electricity, similar to a 600W constant load in my part of the world.

[leptons]

All of my IT equipment in my office is running through a single UPS that measures power consumption.

I do have a bit more than just that server hooked up to it. There's also a Dell i5 running DDWRT as my main gateway/router, the fiber internet modem, a small Synology NAS, a couple of WIFI routers, etc. It all adds up.

That doesn't include my backup server out in the garage with another 8-disk RAID10 array and an LTO tape drive that is often backing up data, 5 more WIFI routers around the property, and 10 or so security cameras. So I'm probably well over $100/mo for all my tech stuff.

[acidburnNSA]

Ryzen 5950x cpu, 64 gb ecc ram, dual 16 tb drives for zfs, Nvidia 5070 gpu.

Way way overspeced for what I listed, but I use it for lots of video processing, numerical simulations, and some local AI too.

I have a similar subset of this stuff running at my mom's house on a 16 GB ram Beelink minicomputer. With openvino frigate can still do fully local object detection on the security case, whish is sweet.

[drnick1]

Not impressive at all. I run just about as many services, plus several game servers, on a Ryzen 5, and most of the time CPU usage is in the low single digits. Most stuff is idle most of the time. Something like a Home Assistant instance used by a single household is basically costless to run in terms of CPU.

[pajamasam]

Not costless in terms of RAM though, surely?

[embedding-shape]

Ultimately, basically. I have two servers in my homelab, one that is more beefy, which hosts a bunch of stuff (basically everything parent outlined + ), including a DHT crawler, download clients, indexers, databases and a lot more. It's sitting and using 16GB (out of available 126GB) right now. Then I have another which only runs the security system + Frigate + Home Assistant, it's using 2.3GB out of 32GB available.

[drnick1]

Web apps like Home Assistant are very light, things like game servers are heavier since they have to load maps etc.

[cyberpunk]

You could easily run all of that on a rpi…

[tclancy]

No, you definitely can’t. Or at least, not 3B+. I wound up buying https://www.amazon.com/ACEMAGICIAN-M1-Computers-Computer-3-2... which was $50 less a month ago (!!) because so many things don’t fit well. Immich is amazing, but you wouldn’t get a lot of the coolness of it if you can’t run the ai bits, which are quite heavy.