Submitted 1 day ago by kiol@lemmy.world to selfhosted@lemmy.world
Curious to know what the experiences are for those who are sticking to bare metal. Would like to better understand who keeps such admins avoiding containers, Docker, Podman, Virtual Machines, etc. What keeps you on bare metal in 2025?
I use Raspberry Pi 4 with 16GB SD-card. I simply don’t have enough memory and CPU power for 15 separate database containers for every service witch I want to use.
Databases on sd cards are a nightmare for sd card lifetimes. I would really recommend getting at least a USD SSD stick instead if you want to keep it compact.
Your SD card will die suddenly someday in the near future otherwise.
Thank you for your advice. I do use a external hard drive for my data.
So, are you running 15 services on the Pi 4 without containers?
I run my NAS and Home Assistant on bare metal.
Both of those are much easier to manage on bare metal. Everything else runs virtualized on my Proxmox cluster, whether it’s Docker stacks on a dedicated VM, an application that I want to run separately in an LXC, or something heavier in its own VM.
Are you concerned about your self-hosted bare metal machine being a single point of failure? Or, are you concerned it will be difficult to reproduce?
Considering I have a full backup, all services are Arch packages and all important data is on its own drive, I’m not concerned about anything
Anything you want dedicated performance on or require fine tuning for a specific performance use cases. Theyre out there.
Erm. I’d just say there’s no benefit in adding layers just for the sake of it.
It’s just different needs. Say I have a machine where I run a dedicated database on, I’d install it just like that because as said there’s no advantage in making it more complicated.
What are you doing running your vms on bare metal? Time is a flat circle.
for work I have a cloud dev VM, in which I run WSL2. so there’s at least two levels of VMs happening, maybe three honestly.
Your phrasing of the question implies a poor understanding. There’s nothing preventing you from running containers on bare metal.
My colo setup is a mix of classical and podman systemd units running on bare metal, combined with a little nginx for the domain and tls termination.
I think you’re actually asking why folks would use bare metal instead of cloud and here’s the truth. You’re paying for that resiliency even if you don’t need it which means that renting the cloud stuff is incredibly expensive. Most people can probably get away with a$10 vps, but the aws meme of needing 5 app servers, an rds and a load balancer to run WordPress has rotted people. My server that I paid a few grand for on eBay would cost me about as much monthly to rent from aws. I’ve stuffed it full of flash with enough redundancy to lose half of it before going into colo for replacement. I paid a bit upfront but I am set on capacity for another half decade plus, my costs are otherwise fixed.
Your phrasing of the question implies poor understanding.
Your phrasing of the answer implies poor understanding. The question was why bare metal vs containers/VMs.
I use k3s and enjoy benefits like the following over bare metal:
These are just the benefits just for one server. Add more and the benefits increase.
In my case it’s performance and sheer RAM usage.
GLM 4.5 needs like 112GB RAM and absolutely every megabyte of VRAM from the GPU. It simply cannot afford the overhead. I think containers may slow down CPU<->GPU transfers slightly, but don’t quote me on that.
Can anyone confirm if containers would actually impact CPU to GPU transfers
To be clear, VMs absolutely have overhead but Docker/Podman is the question. It might be negligible.
And this is a particularly weird scenario (since prompt processing literally has to shuffle 112GB over the PCIe bus for each batch). Most GPGPU apps aren’t so sensitive to that.
I don’t host on containers because I used to do OS security for a while.
Say more, what did that experience teach you? And, what would you do instead?
@kiol I mean, I use both. If something has a Debian package and is well-maintained, I'll happily use that. For example, prosody is packaged nicely, there's no need for a container there. I also don't want to upgrade to the latest version all the time. Or Dovecot, which just had a nasty cache bug in the latest version that allows people to view other peoples' mailboxes. Since I'm still on Debian 12 on my mail server, I remain unaffected and I can let the bugs be shaken out before I upgrade.
@kiol On the other hand, for doing builds (debian packages and random other stuff), I'll use podman containers. I've got a self-built build environment that I trust (debootstrap'd), and it's pretty simple to create a new build env container for some package, and wipe it when it gets too messy over time and create a new one. And for building larger packages I've got ccache, which doesn't get wiped by each different build; I've got multiple chromium build containers w/ ccache, llvm build env, etc
@kiol And then there's the stuff that's not packaged in Debian, like navidrome. I use a container for that for simplicity, and because if it breaks it's not a big deal - temporary downtime of email is bad, temporary downtime of my streaming flac server means I just re-listen to the stuff that my subsonic clients have cached locally.
I have a single micro itx htpc/media server/nas in my bedroom. Why use containers?
Do you back it up?
My raid is rsync :)
I would always run proxmox to set up docker VMs.
I found Talos Linux, which is a dedicated distro for kubernetes. Which aligned with my desire to learn k8s.
It was great. I ran it as bare-metal on a 3 node cluster. I learned a lot, I got my project complete, everything went fine.
I will use Talos Linux again.
However next time, I’m running proxmox with 2 VMs per node - 3 talos control VMs and 3 talos worker VMs.
I imagine running 6 servers with Talos is the way to go. Running them hyperconverged was a massive pain. Separating control plane and data/worker plane (or whatever it is) makes sense - it’s the way k8s is designed.
It wasn’t the hardware that had issues, but various workloads. And being able to restart or wipe a control node or a worker node would’ve made things so much easier.
Also, why wouldn’t I run proxmox?
Overhead is minimal, get nice overview, get a nice UI, and I get snapshots and backups
What hardware are you running it on?
3x minisforums MS-01
I thought about running something like proxmox, but everything is too pooled, too specialized, or proxmox doesn’t provide the packages I want to use.
Just went with arch as the host OS and firejail or lxc any processes i want contained.
I’ve never installed a package on proxmox.
I’ve BARELY interacted with CLI on proxmox (I have a script that creates a nice Debian VM template, and occasionally having to really kill a VM).
What would you install on proxmox?!
Firmware update utilities, host OS file system encryption packages, HBA management tools, temperature monitoring, and then a lot of the packages had bugs that were resolved with newer versions, but proxmox only provided old versions.
For me it’s lack of understanding usually. I haven’t sat down and really learned what docker is/does. And when I tried to use it once I ended up with errors (thankfully they all seemed contained by the docker) but I just haven’t gotten around to looking more into than seeing suggestions to install say Pihole in it. Pretty sure I installed Pihole outside of one. Jellyfin outside, copyparty outside, and I something else im forgetting at the moment.
I was thinking of installing a chat app in one, but I put off that project because I got busy at work and it’s not something I normally use.
I guess I just haven’t been forced to see the upsides yet. But am always wanting to learn
containerisation is to applications as virtual machines are to hardware.
VMs share the same CPU, memory, and storage on the same host.
Containers share the same binaries in an OS.
When you say binaries do you mean locally stored directories kind of like what Lutris or Steam would do for a Windows game. (Create a false c:\ )
TrueNAS is on bare metal has I have a dedicated NAS machine that’s not doing everything else and also is not recommended to virtualize. Not sure if that counts.
Same for the firewall (opnsense) since it is it’s own machine.
Have you tried running containers on Truenas?
No because I run my containers elsewhere, not on the NAS
I generally abstract to docker anything I don’t want to bother with and just have it work.
If I’m working on something that requires lots of back and forth syncing between host and container, I’ll run that on bare metal and have it talk to things in docker.
Ie: working on an app or a website or something in language of choice on framework of choice, but postgres and redis are living in docker. Just the app I’m messing with and it’s direct dependencies run outside.
The pi 3. You can’t exactly put containers into potatoes, can you.
I mean, I did
I’m selfhosting Forgejo and i don’t really see the benefit of migrating to a container, i can easily install and update it via the package manager so what benefit does containerization give?
If you don’t exceed a single deployment, it’s fine.
If however in the future, you want to add additional services to your host, let’s say an alerting or status system, it’s a lot easier to declare everything in a single place and then attach a reverse proxy to manage networking multiple services on one host.
Depends on the application for me. For Mastodon, I want to allow 12K character posts, more than 4 poll question choices, and custom themes. Can’t do it with Docker containers. For Peertube, Mobilizon, and Peertube, I use Docker containers.
Why could you not have that Mastodon setup in containers? Sounds normal afaik
I’ll chime in: simplicity. It’s much easier to keep a few patches that apply to local OS builds: I use Nix, so my Mastodon microVM config just has an extra patch line. If there’s a new Mastodon update, the patch most probably will work for it too.
Yes, I could build my own Docker container, but you can’t easily build it with a patch (for Mastodon specifically, you need to patch js pre-minification). It’s doable, but it’s quite annoying. And then you need to keep track of upstream and update your Dockerfile with new versions.
I’m running a TrueNAS server on bare metal with a handful of hard drives. I have virtualized it in the past, but meh, I’m also using TrueNAS’s internal features to host a jellyfin server and a couple of other easy to deploy containers.
So Truenas itself is running your containers?
Yeah, the more recent versions basically have a form of Docker as part of its setup.
I believe it’s now running on Debian instead of free BSD, which probably simplified the containers set up.
I’m doing this on a couple of machines. Only running NFS, Plex (looking at a Jellyfin migration soon), Home Assistant, LibreNMS and some really small other stuff. Not using VMs or LXC due to low-end hardware (pi and older tiny pc). Not using containers due to lack of experience with it and a little discomfort with the central daemon model of Docker, running containers built by people I don’t know.
The migration path I’m working on for myself is changing to Podman quadlets for rootless, more isolation between containers, and the benefits of management and updates via Systemd. So far my testing for that migration has been slow due to other projects. I’ll probably get it rolling on Debian 13 soon.
my two bare metal servers are the file server and music server. I have other services in a pi cluster.
file server because I can’t think of why I would need to use a container.
the music software is proprietary and requires additional complications to get it to work properly…or at all, in a container. it also does not like sharing resources and is CPU heavy when playing to multiple sources.
if either of these machines die, a temporary replacement can be sourced very easily(e.g. the back of my server closet) and recreated from backups while I purchase new or fix/rebuild the broken one.
IMO the only reliable method for containers is a cluster because if you’re running several containers on a device and it fails you’ve lost several services.
Cool, care to share more specifics on your Pi Cluster
I followed one of the many guides for installing proxmox on Rpis. 3node, 4gb rpi4s
I use the cluster for lighter services like Trilium, FreshRss, secondary DNS, a jumpbox… and something else I forget. I’m going to try immich and see how it performs.
my recent goto for cheap($200-300) servers are Debian + old Intel Macbook pros. I have two Minecraft bedrock servers on MBPs… one an i5, the other an i7.
I also use a Lenovo laptop to host some industrial control software for work.
Not knowing about Incus (LXD). It’s a life changer. Would never run any service on bare metal again.
Using GenAI to develop my Terraform and Ansible playbooks is magical. Also, use it to document everything in beautiful HTML docs from the outputs. Amazing.
After many failures, I eventually landed on OMV + Docker. It has a plugin that puts the Docker management into a web UI and for the few simple services I need, it’s very straightforward to maintain. I don’t cloud host because I want complete control of my data and I keep an automatic incremental backup alongside a physically disconnected one that I manually update.
Cool, how are you managing your disks? Are you overall happy with OMV?
Very happy with OMV. It’s not crazy customizable, so if you have something specialized, you might run into quirks trying to stick to the Web UI, but it’s just Debian under the hood, so it’s pretty manageable. 4x1TB drives RAID 5 for media/critical data, OS drive, and a Service data drive (databases, etc). Then an external 4TB for the incremental and another external 4TB for the disconnected backup.
Depends on the application. My NAS is bare metal. That box does exactly one thing and one thing only, and it’s something that is trivial to setup and maintain.
Nextcloud is running in docker (AIO image) on bare metal (Proxmox OS) to balance performance with ease of maintenance. Backups go to the NAS.
Everything else is running on in a VM which makes backups and restores simpler for me.
hperrin@lemmy.ca 17 hours ago
There’s one thing I’m hosting on bare metal, a WebDAV server. I’m running it on the host because it uses PAM for authentication, and that doesn’t work in a container.