this post was submitted on 25 May 2026
35 points (97.3% liked)

Selfhosted

59631 readers
611 users here now

A place to share alternatives to popular online services that can be self-hosted without giving up privacy or locking you into a service you don't control.

Rules:

  1. Be civil: we're here to support and learn from one another. Insults won't be tolerated. Flame wars are frowned upon.

  2. No spam posting.

  3. Posts have to be centered around self-hosting. There are other communities for discussing hardware or home computing. If it's not obvious why your post topic revolves around selfhosting, please include details to make it clear.

  4. Don't duplicate the full text of your blog or github here. Just post the link for folks to click.

  5. Submission headline should match the article title (don’t cherry-pick information from the title to fit your agenda).

  6. No trolling.

  7. No low-effort posts. This is subjective and will largely be determined by the community member reports.

Resources:

Any issues on the community? Report it using the report flag.

Questions? DM the mods!

founded 3 years ago
MODERATORS
ruud@lemmy.world
Loki@lemmy.world
CannaVet@lemmy.world
devve@lemmy.world
HybridSarcasm@lemmy.world
HybridSarcasm@lemmy.hybridsarcasm.xyz
35
Looking to build a K8s cluster, what are your strategies for finding affordable hardware? (lemmy.world)
submitted 1 week ago by Eldaroth@lemmy.world to c/selfhosted@lemmy.world
43 comments fedilink hide all child comments
 

I currently run all my self-hosted apps either on Podman in a VM or in LXCs on Proxmox. For hardware, I'm using a Chinese-made mini computer with an Intel N150 and 16GB of DDR5 RAM that I bought before the whole AI hype started. I also have a Synology NAS that I use mainly for media and photo storage.

I've been thinking about tinkering with Kubernetes in my homelab for a while now (I already use it extensively at work, so I'm quite familiar with it), and I started looking around for used hardware to use as bare metal nodes. Nothing fancy—I'm looking for 1 or 2 mini servers or SFF with at least 16GB of memory and a decent CPU (4–6 cores). But with current prices, even decently priced used hardware (~200–250€) is quite difficult to find in Europe, and most of it is HP stuff with Lenovo being a rare breed around here. I won't even get started on newly bought hardware...

If you've bought hardware in this market recently, how did it go for you? Or are most of you holding out for now, waiting for better times?

top 43 comments
sorted by: hot top controversial new old
[–] moonpiedumplings@programming.dev 10 points 1 week ago (2 children)

Even used stuff is expensive nowadays.

Anyway, you can buy these used refurbished small form factor business PC's.

These things: https://www.servethehome.com/introducing-project-tinyminimicro-home-lab-revolution/

I was recently at a tech conference and I met a guy who was selling 16 gb ram one's for 30 usd, since they had managed to track them down that cheap in bulk somewhere. What happens is that corporations or govt get rid of them due to warranty expiry, so they need to be offloaded somewhere.

You probably won't be able to find them that cheap but it's definitely more affordable than new stuff. It used to be cheaper but I mostly see 200 usd in my searches.

[–] Eldaroth@lemmy.world 2 points 1 week ago

Thanks for the link! I think I saw this project mentioned in some other post as well, but honestly didn't look further into it then.

Oh wow 30 bucks a unit for 16 GB, that's just insanely lucky to find a deal like that.

Looking around in my region, there is from time to time the odd deal where you could get for example a HP EliteDesk 800 G4 Mini with 16 gigs for like below 200€ if you're lucky and with the right timing. But usually they will net you 260€ upwards.

[–] irmadlad@lemmy.world 2 points 1 week ago (2 children)

Even used stuff is expensive nowadays.

No shit! I was shopping NewEgg for some used drives. Found a 500 GB HDD for $36 USD which I immediately clicked on only to find this warning:

Used - Very Good Erased (DoD 5220.22 M Compliant) and tested. There may be drives that have up to 25 bad sectors. There may be writing or >markings on the drive or the label may vary slightly from picture.

WTF. Who would buy a used HDD with up to 25 bad sectors?

[–] Andres4NY@social.ridetrans.it 3 points 1 week ago (1 children)

@irmadlad @moonpiedumplings Enterprise drives w/ bad sectors and a 5yr warranty? Me (raises hand)!

With snapraid and backups and the warranty, I'm perfectly happy to use enterprise drives _where the drive errors aren't increasing_. Worst case, I lose some of those linux distributions and have to re-download them. I wouldn't use one without a warranty though (and I certainly wouldn't waste my time on a 500gb drive, the last hdd I bought w/ sector errors was 14TB for $140).

[–] irmadlad@lemmy.world 1 points 1 week ago* (last edited 1 week ago) (2 children)

14TB for $140).

Link me up, Scotty

(raises hand)!

Really. I would have never guessed. I mean, I've always viewed bad sectors as a sign of a failing drive. Just a really bad day in the future, marked as an appointment.

What kind of millage on the average do you get out of something like that?

[–] moonpiedumplings@programming.dev 4 points 1 week ago (1 children)

When people say that linux adds a second life to bad hardware, they don't just mean making 8 gb of ram usable again. They also mean stuff like this, using dying hardware to it's last breath.

Related: https://lorenz.brun.one/dealing-with-bad-ram-on-linux/

You can mark the failing parts of ram as explicitly bad so Linux avoids them, just like with hard drives. Another way to get more lifespan out of this hardware.

[–] irmadlad@lemmy.world 2 points 1 week ago

You can mark the failing parts of ram as explicitly bad so Linux avoids them, just like with hard drives. Another way to get more lifespan out of this hardware.

I love this. I am being schooled. I honestly have never heard of what you speak of. But by golly I'm going to spool up on the topic.

[–] Andres4NY@social.ridetrans.it 2 points 1 week ago (1 children)

@irmadlad I can't tell you averages (I'm not running massive storage servers with hundreds of drives here), but I ordered this particular one in Oct 2025. It's still going strong. There's another one that I bought back in 2019 or 2020 that I used for a few years before replacing it due to needing more space. Meanwhile, I've used plenty of consumer drives over the years that were a lot less reliable, so I have different rules for consumer drives (toss 'em at the first error) vs enterprise drives.

An order confirmation email from GoHardDrive (10/1/25), showing a purchase of a Western Digital Ultrastar 14TB 7200RPM drive (refurbished - grade B) w/ 5yr warranty. For $139.99.

[–] Andres4NY@social.ridetrans.it 2 points 1 week ago

@irmadlad Also, I would think that companies offering long warranties on refurbished drives are playing the odds in a way that makes them money. It probably wouldn't be profitable if they sold drives w/ bad/reallocated sectors and the majority of them died within the warranty period.

My assumption is that all drives will die (or suffer corruption) at the worst possible time, so do proper backup/scrubbing. Then look for deals where I can.

[–] adarza@lemmy.ca 2 points 1 week ago (1 children)

i have no idea.

i personally bin (recycle bin, that is) any drive (ssd or hdd) that has any abnormality in its internal diagnostics data, or won't pass short and long diagnostics tests even if their stats are clean.

[–] irmadlad@lemmy.world 1 points 1 week ago* (last edited 1 week ago)

I've heard this dude named @Andres4NY@social.ridetrans.it he uses them. I've never heard of that. TIL

[–] irmadlad@lemmy.world 5 points 1 week ago

Nothing fancy—I’m looking for 1 or 2 mini servers or SFF with at least 16GB of memory and a decent CPU (4–6 cores)

Not in the EU, so......

The Optiplex 7020 SFF with the i7-4790 chip are pretty robust little computers, regularly sell on Amazon for less than $200. The 9020 MT is a bit older, however still capable @ well under $200 USD on Amazon. I'd upgrade the 9020 to a i7-4790 chip which you can normally find, used, on ebay for less than $50 USD. The models in that range take DDR3 which you can find good prices for at https://www.memorystock.com/. I've been using them for years. The bad part about shopping for used computers on ebay is that, while you might be able to find some killer deals initially, most often they stick it to you good with shipping.

Also: Dell OptiPlex 3050 Micro Not too shabby either.

I just recently got a Optiplex 9020 MT 16 GB RAM for $124 USD / free shipping off of Amazon. Shopped around ebay for the i7-4790 which cost me $35 USD, and 16 GB DDR3 RAM for $50. Not a bad little set up for right at the $200 USD mark. I'm setting it up for a friend of mine's son who wants to get into selfhosting at the age of 10. It's no threadripper but a solid piece.

[–] zzffyfajzkzhnsweqm@sh.itjust.works 3 points 1 week ago (1 children)

Do not run kubernetes nodes as LXC. I mean you can for testing etc... BUt you will have to break sandboxing because of networking capabilities kubernetes requires. At least that was the case for K3S. You could probably have better experience with Kind.

[–] Eldaroth@lemmy.world 2 points 1 week ago (1 children)

Not planning to run nodes as LXC at all. Just using them in my current setup for running services like Wireguard or Jellyfin.

At most I would create VMs for nodes, but still on the fence if I should just go bare metal or proxmox with virtualized nodes (VM).

[–] zzffyfajzkzhnsweqm@sh.itjust.works 1 points 1 week ago

I would go with two (or more) proxmox nodes working as one. And then running even more VMs on them as kubernetes nodes. Proxmox handles lots of stuff for yout that is harder to handle in phusical machine.

[–] Routhinator@startrek.website 3 points 1 week ago (1 children)

I used Zimaboards. They are affordable and all you need if you couple the cluster with a TrueNAS Scale box and setup Democratic CSI.

[–] Eldaroth@lemmy.world 1 points 1 week ago (1 children)

Like the original Zimaboard? How do you like it? Reading online in forums or reddit, the feedback is really mixed.

I was looking at the Zimaboard 2, but the board alone is like 300€+ and then buying DDR5 with current prices...

[–] Routhinator@startrek.website 1 points 1 week ago* (last edited 1 week ago)

Zimaboards do not need RAM. They include it (do not confuse it with the Zimablade).

I just use the default hardware with the Zimaboards. Zimaboard 832s for the masters and nodes, and I recently added a couple of Zimaboard 2s for larger nodes as those have 16gb.

I like the boards, they perform well and the intel cpus in them have internal GPUs so for light transcoding or GPU loads they can be leveraged without need for GPU cards. After trying several other x86 based options, Zimaboards are my preferred option.

[–] PseudoRandomGermanPerson@programming.dev 3 points 1 week ago

My janky(-ish) but fun plan is to use laptop motherboards, either from broken laptops or AliExpress. Wire them up with an ATX powersupply and that should be a cheap and powerful cluster. If you want something less hacky the same idea but with used mini-PCs from offices.

[–] Reannlegge@lemmy.ca 3 points 1 week ago (1 children)

I personally have a handful of pi’s doing things, with plans to get more to do more (I have one in the mail to become my OpenWRT router). I was thinking of getting something simple to be able to do STT and TTS as I want to replace my HomePods as I slowly leave Apple and have have no interest in joining Amazon or Google for home automation, so I looked at my university’s used computers. I am not really a fan of buying some random persons old hardware but maybe a business is selling their old stuff.

[–] Eldaroth@lemmy.world 1 points 1 week ago (1 children)

I assume you are using pi 4 and/or 5? I would have some 3b+ laying around which don't see any use nowadays. But can't quite imagine to build a cluster only made up of these, I think they are too underpowered, unless being used as like a small additional worker node.

And buying new Pi 5 with 16GB of memory would set me back at least 200€ a unit, not including power supply or any case.

[–] paulcdb@lemmy.world 1 points 1 week ago (1 children)

I’m also in the PI 4 group. Its worth watching ebay, etc for people selling pi clusters. I picked up my 8 node 4GB pi cluster fairly cheap.

Also if you’re learning and find yourself struggling with yaml files, I found Claude helped a lot to get running. Create the file, dump it into Claude and let it re-output the file. Give it the error log helps if it gets stuck.

Also if you explain your cluster setup, I built mine with Talos Linux so setup VIP, etc and it’s pretty good at generating a working setup.

I’m not a huge fan of AI but I struggled with yaml formatting so it might suck for a lot of stuff but formatting stuff properly is great.. that and scanning logs for errors! 😎

[–] Eldaroth@lemmy.world 1 points 1 week ago

I liked my Pi 3B+ for tinkering around and run some services like vaultwarden or a VPN node with PiVPN (first OpenVPN, later Wireguard). They are great for stuff like that and don't use much power. Not sure it makes sense to add them as nodes though as 2GB is stretching it, although Talos OS has a minimal requirement for worker nodes of 1GB, so I might still try 😄

As you can see I plan to use Talos as well, it sounds really promising what I read about it. For a CNI I plan to use Cilium and for distributed storage I might try Longhorn and see how it works. Apart from that I am still on the fence if I should run my nodes bare model or put a virt layer in between with Proxmox.

[–] slazer2au@lemmy.world 3 points 1 week ago (2 children)

Honest question, why are you looking to buy hardware to do this when you can mess with K8S in VMs on your existing proxmox host?

[–] Eldaroth@lemmy.world 2 points 1 week ago (2 children)

Fair point. If I don't want to just run a single node cluster, I will need at least 2-3 nodes aka VMs. Looking at like 2-4GB (as a bare minimum imho) memory for each, I would already eat into like half of my current 16GB of available memory. And this mini server is already running a lot of my critical apps, using more than half of the memory.

[–] moonpiedumplings@programming.dev 2 points 1 week ago* (last edited 1 week ago) (2 children)

I run a single node cluster.

My single node has 256 gb of ram and 24 cores. I do this because, if you want a lot of ram/cores/storage, it is cheaper to get a used "tower server" type device and then upgrade it as you go over time, than it is to buy entirely new devices for every bit of ram you want to add to the cluster.

I like kubernetes because I like configuration as code, gitops, the way it abstracts over components so I can swap components out easily, the way that helm charts are an easier way of orchestrating containers, and a bunch of other things.

Clustering is merely one of many benefits of kubernetes, one that isn't particularly important to me. Although, my opinion on that has changed somewhat recently. Waiting for a reboot is annoying, since I am rebooting the whole thing and I have to wait for each service to go down or come up before the machine reboots properly. But if I was running kubernetes as a virtual machines inside incus with multiple nodes, I could update each node one by one without the whole thing going down. Or, I could snapshot them, allowing me to reboot the host without waiting for kubernetes. But these things are mostly just somewhat nice to have, rather than a core feature I really require.

[–] Eldaroth@lemmy.world 3 points 1 week ago

Oh wow okay, if I'd go down that route I would definitely do multiple VMs on that host. In my opinion, the whole clustering and self healing / HA aspect of Kubernetes is why I want to switch to it. I can do gitops with Podman/Docker as well, in fact I already do that including a renovate pipeline on my sel-hosted Forgejo instance. But having the redundancy of several nodes, if one goes down the service will still work or at least will be re-deployed within a couple of seconds, provided there is distributed storage (longhorn, ceph or even nfs cis) if its a stateful app.

[–] hirihit640@sh.itjust.works 1 points 1 week ago (1 children)

have you found many self-hosted services that suppprt that kind of HA? I can't imagine services like torrent clients allowing you to stream writes to one node while replicating to the other, though maybe I'm misunderstanding the setup

[–] moonpiedumplings@programming.dev 0 points 1 week ago (1 children)

They don't have to support it. It's more that the database they are connecting to supports HA, and/or both are using shared storage. So when one container dies, kubernetes restarts another container on another node with the same shared storage attached.

Docker is often configured to automatically restart containers when a container dies, just in case it's a one off bug or something like that, and kubernetes is like a more resilient version of that.

[–] hirihit640@sh.itjust.works 1 points 1 week ago (1 children)

I'm aware of databases that support HA, but the vast majority of self-hosted apps I've encountered use file storage, even if they have a database as well. It sounds like you're proposing shared storage like an NFS share. But if you're upgrading nodes, at some point you have to upgrade the node hosting the shared storage right? Wouldn't that take down all services? Unless you use a distributed storage system, but I've heard those can get very complicated...

[–] moonpiedumplings@programming.dev 1 points 1 week ago (1 children)

Kubernetes makes distributed storage easy.

Basically, all the components get deployed for you, since that's part of what kubernetes is good at.

And then, services/containers can provision storage by requesting storage via making a "claim" and whatever distributed storage providee you have gives it to you.

[–] hirihit640@sh.itjust.works 1 points 1 week ago (1 children)

Have you tried using Ceph or other distributed storage systems in your kubernetes cluster?

[–] moonpiedumplings@programming.dev 0 points 1 week ago (1 children)

Well, I run a one node cluster...

But yes, I did use ceph via rook-ceph, because Openstack (a locally hosted AWS alternative), at least the Kubernetes version, wanted a ceph "cluster" to store stuff on.

Longhorn is much easier. Although again, my "cluster" was one node. I deployed it because I wanted snapshots.

[–] hirihit640@sh.itjust.works 1 points 6 days ago (1 children)

Wow cool! I believe you're the first person I've met that actually used a cluster FS (in their homelab at least). I looked into it myself but it felt like nobody was really using it so I didn't bother.

Does it involve much more work or is it a fairly transparent replacement to traditional storage options? Assuming one is already using Kubernetes. I'm wondering if it's worth it to switch to a cluster FS for everything, like Radicale or Tiddlywiki.

[–] moonpiedumplings@programming.dev 1 points 6 days ago

On kubernetes it's pretty much the same amount of work. Every possible storage option exposes a generalized, abstracted "storageclass", from which storage can be provisioned and mounted into containers.

https://kubernetes.io/docs/concepts/storage/storage-classes/

[–] slazer2au@lemmy.world 1 points 1 week ago

That's fair.

[–] jimmy90@lemmy.world 1 points 1 week ago (1 children)

I tried this on proxmox and ran into a lot of installation issues. Any tips?

[–] slazer2au@lemmy.world 1 points 1 week ago (1 children)

Have you watched the Learn Linux TV video series on YouTube? He does a good job at walking you through k8s from nothing using cloud init images, but you don't need that part.

[–] jimmy90@lemmy.world 2 points 1 week ago

Will check it out thanks

[–] utjebe@reddthat.com 2 points 1 week ago

€200-250 is the entry point for something like 10th gen i5. You mostly get 1 year warranty on refurbished hw and as long as it is not DoA, it will be fine. These machiness are for sale because of accounting, not because of being broken

[–] Demivan@lemmy.world 2 points 1 week ago

Non hardware answer: but for bare metal Kubernetes install I really recommend Talos OS. It is easy to install, secure and very lightweight. So you can squeeze more out of your hardware.

[–] Decronym@lemmy.decronym.xyz 2 points 1 week ago* (last edited 6 days ago)

Acronyms, initialisms, abbreviations, contractions, and other phrases which expand to something larger, that I've seen in this thread:

Fewer Letters More Letters
DNS Domain Name Service/System
HA Home Assistant automation software
~ High Availability
LXC Linux Containers
NFS Network File System, a Unix-based file-sharing protocol known for performance and efficiency
VPN Virtual Private Network
k8s Kubernetes container management package

6 acronyms in this thread; the most compressed thread commented on today has 9 acronyms.

[Thread #316 for this comm, first seen 26th May 2026, 10:20] [FAQ] [Full list] [Contact] [Source code]

[–] minfapper@piefed.social 1 points 1 week ago

Dunno about Europe, but if anyone is reading this from America, here you go:

https://www.ebay.com/sch/i.html?_nkw=dell+optiplex&_sacat=0&_sop=15&rt=nc&LH_BIN=1

They are a little more expensive than the last time I looked, but still cheaper than raspberry pi.