I am several months into the self-hosting journey and I feel I have outgrown my Pi 4 B 8GB. I’m only running around 3 dozen containerized services and it seems to struggle to keep up. But I’m not sure of the best bang for my buck. I’d like good, long-term performance, but I don’t really have a grand lying around for a Lenovo Tiny or Dell Optiplex or ASUS NUC. I’m thinking of buying an SSD to boot from, but will this even help much? For $350-500, could I make a more cost effective homeserver upgrade?
A grand?? You can pick up a used Lenovo Tiny for 50 bucks (US) on EBay.
This is what i did. In europe, viable options start at 200€ on ebay (imo). If your use case outgrows one lenovo tiny (which is unlikely since you’re coming from a pi), you can buy more / other tiny pcs / a desktop pc / a server rack and put proxmox on everything for running services inside a cluster.
You can get a decent Lenovo Tiny for 75€ + 10€ IMO. On eBay.
Like some 2-4 core 4-8GB 120-240GB SSD. Blasts a hole in any SOC like the raspberry.
Wow, that’s pricey!
Just get a used ultra-small form factor PC a la the Tiny, Mini, or Micro series. A higher-end one which is 7 generations old will still absolutely destroy the Pi in terms of performance.
Once I gave up (for now) on doing all this on ARM and switched back to x86, everything got way easier to actually accomplish.
You could build a Framework desktop with one of their past gen motherboards. Also look for surplus servers. The first 1U servers I bought were only about $150 and lasted many years.
Check out minisforum, for example this intel mini-pc. They have a ton of selection, not just that one example.
This seems like exactly the site I’m looking for! Thanks
You also have beelink that makes these small PCs
For $350-500 you could easily get a used desktop and processor with 16-32 gb ddr4. But it sort of depends on your home lab goals and workloads. Do you need a lot of storage? Are you CPU bound or memory bound? Some people will suggest used Dell/HP servers, and they’ll look affordable, but keep in mind enterprise gear will eat power and is usually loud. Personally I’d go for a used AMD 5800 or 5900 processor and mobo, install your favorite Linux, and call it a day. AMD processors don’t have quick sync which makes them slightly worse for plex hosting but better for everything else.
Not sure if I’m CPU-bound or RAM-bound, but I’d hazard to guess both lol And the lower the power consumption the better as this is an always-on, very passive deal.
Personally, I only plan to run another dozen or so containers. What I want most is reliability/stability. I want everything I set up to, once it works, continue working without issue. This is where the Pi has begun frustrating me as it seems to just seize up sometimes. I do need as much storage as possible as I’m a bit of a media hoarder, but that can always be solved with more drives.
Makes sense. I think you’d be fine with pretty much any modern(post DDR4) motherboard/CPU combo these days. I feel Linux hardware support is only really shakey if you’re using a SoC without upstream patches or if you’re using brand new hardware/laptops. With that being said if you’re running a lot of containers on one host have you looked into docker compose or kubernetes(k8s)? Maybe k8s is overkill for home use, but both offer support to restart containers if a health check fails. With k8s you also can spread out containers across multiple physical node, so you could just add a second RPI and “double” your resources.
Also worth looking at k3s if its running on a single node to reduce resource usage
A second Pi? According to the RPI Locator, Raspberry Pi’s are available for list price now.
My use case is a bit different in that I don’t know what containerized services I want to run, except that I want to play with Kubernetes. Raspberry Pi still seems like a good choice and I may restart that project soon
I’d rather keep all the services to a single more powerful device and then relegate the Pi to more specialized, Pi-related tasks like a smart doorbell cam or Home Assistant Hub.
Cost of electricity is non zero. Distributed computing between pis might be the most cost effective way (hardware and electricity)
You could pick up a used laptop for pretty cheap. Low TDP, leagues ahead of a Pi
This is a great suggestion. There’s so much e-waste out there that could be more than powerful enough to be a major upgrade from a Pi. If OP has a PC built in the last 15 years it’s almost certainly the cheapest (and greenest) solution.
As the computer nerd in my friends/family circle, Im constantly being gifted old PCs from people. Some I keep and repurpose into servers/other things and others I just recycle for them. I also am an avid PC gamer so when my rig gets retired I usually upgrade the other servers internals from parts I had left over. Don’t know if something else is viable for you to keep costs down, even if you can get the PC without a HDD but everything else.
For me, the small energy footprint of an ARM machine is really important for home usage so I personally went on multiple occasions with Odroid’s offerings and as long as you have a tinkerer’s soul, I can really recommend it.
In that case, I can recommend minicomputer’s like HP EliteDesk G2 800 Mini. You can get them with a variety of intel CPUs, they can take up to 32GB RAM, they have slot for M.2 disks and a regular 2.5" SSD – and they hardly use any power when idle, between 5 to 10 watts, depending on the CPU and CPU governor settings. They are sold used for ~€50 and if you buy newer generations you’ll get even more umpfh for a bit more cash.
In other words, very competetive with the Pi’s, only more available, cheaper and about the same power consumption!
And look better.
I also want to prioritize power consumption just because I can’t afford server rack levels of electricity, so I will have to check that out.
Intel’s low power offerings are sometimes even less power hungry than a RPi and handle more stuff. I like Asrock’s line of CPU-onboard motherboards and use one myself. You get the convenience of a full x86 machine but it sips power. Mine peaks at ~36W with full load on CPU, GPU, RAM and 4 SSDs or disks. Usually it is much much lower. You can always go smaller with an Atom x5 z8300 (~2W Idle without disks or network, 6W with both and some load), but those are getting a little old and newer stuff is better and more feature-rich. Maybe an N100 machine with 4 or 8 gigs of RAM are a good option for you? Don’t go overboard with RAM if you are using docker for everything anyways. I use 8 but 4 would be more than enough for me and my countless containers. I run Nextcloud, Jellyfin, Paperless-ngx, Resilio, Photoprism and a few more. Only the minecraft server benefits from more than four. Very happy with my J5005 board.
I got a free computer and upgraded the processor to an i7-6700T (eBay) and some old SSDs. It measured around 15W and I haven’t had any problems with it. It is miles ahead of using any Pi or ARM-based SBC. I would really recommend just finding a used computer nearby, if possible. Craigslist and Facebook Marketplace has some killer deals.
If containerized, scale horizontal and load balance.
Could you elaborate on that possibly?
Aka, buy a second Rasberry 😅
Pretty much this. There is a significant learning curve to kubernetes (or similar, I guess?), but it’s a handy skill to be familiar with. If things basically run but with degraded performance on one raspberry, there is no real need for a vertical performance increase. Adding the second will provide almost double capacity, with some HA added for fun.
Acronyms, initialisms, abbreviations, contractions, and other phrases which expand to something larger, that I’ve seen in this thread:
Fewer Letters More Letters HA Home Assistant automation software ~ High Availability LXC Linux Containers MQTT Message Queue Telemetry Transport point-to-point networking NUC Next Unit of Computing brand of Intel small computers PSU Power Supply Unit RPi Raspberry Pi brand of SBC SBC Single-Board Computer SSD Solid State Drive mass storage k8s Kubernetes container management package
[Thread #146 for this sub, first seen 18th Sep 2023, 18:05] [FAQ] [Full list] [Contact] [Source code]
if you’re in the u.s. check walmart for clearance desktops. we just picked up (to replace an older neighbor’s dead atom-based desktop) a new i3, 8gb, 256gb nvme, slim desktop. inside it had mount points and the cables for two sata (3.5in + 2.5in) and a slim optical (or a ‘creatively’ mounted second 2.5in). we were going there to pick it up at a bit over $300, as i had seen it there the previous day. surprise! it had just got marked down again to ~ $200. original price was $400.
Damn that’s a steal! Guess I may have to go the old brick and mortar route.
I recently replaced a pi4 with an intel n95 based mini-pc and it’s been an absolute joy. I moved a few of my services and VMs over from the main server. It runs my Homeassistant via HAOS in a VM significantly faster than the pi4, another VM with access to the IGPU runs a 4k dashboard feed into my video distribution matrix, a few containers for simple things like MQTT and Adguard Home (like pihole) and it has room to do more.
The whole computer with 16gb of ram and 256gb SSD cost about the same as a pi4 8gb did when the shortage was at its worst.
The other option of course, is a cheap older optiplex, for under $200 you can get quad core sky lake or kaby lake generation processors, 16gb of ram and room for a couple of SSDs, a bulk storage HDD and a couple of low profile PCIE cards.
I can vouch for the power of a nuc, they’re basically laptop grade hardware.
Dominant failure modes are fan failure and ssd failure. The latter can be solved by using a quality ssd, the former by keeping your nuc out of dirty areas. You can clean it up if it gets dirty, but it’s a high risk operation, I’ve seen fan blades break.
