I’m still running a 6th-generation Intel CPU (i5-6600k) on my media server, with 64GB of RAM and a Quadro P1000 for the rare 1080p transcoding needs. Windows 10 is still my OS from when it was a gaming PC and I want to switch to Linux. I’m a casual user on my personal machine, as well as with OpenWRT on my network hardware.
Here are the few features I need:
- MergerFS with a RAID option for drive redundancy. I use multiple 12TB drives right now and have my media types separated between each. I’d like to have one pool that I can be flexible with space between each share.
- Docker for *arr/media downloaders/RSS feed reader/various FOSS tools and gizmos.
- I’d like to start working with Home Assistant. Installing with WSL hasn’t worked for me, so switching to Linux seems like the best option for this.
Guides like Perfect Media Server say that Proxmox is better than a traditional distro like Debian/Ubuntu, but I’m concerned about performance on my 6600k. Will LXCs and/or a VM for Docker push my CPU to its limits? Or should I do standard Debian or even OpenMediaVault?
I’m comfortable learning Proxmox and its intricacies, especially if I can move my Windows 10 install into a VM as a failsafe while building a storage pool with new drives.
You must log in or register to comment.
My needs are pretty similar to yours and I’ve recently moved back to using hypervisors after running everything from Debian to Arch to NixOS bare-metal over the last decade or so. It’s so easy to bring-up/tear-down environments, which is great for testing things and pretty much the whole point of a homelab. I’ve got a few VMs + one LXC running on Proxmox with some headroom on a 6th gen i7, you should be fine resource wise tbh. Worth mentioning that you’ll most likely need to passthrough your drives to the guest VM which is not supported via the webUI, but the config is documented on their wiki.
Overall, I’m happy with this setup and loving CoreOS as a base-OS for VMs and rootless podman containers for applications.
use nixos! you won’t regret it
Thanks everyone, I feel much better about moving forward. I’m leaning towards Proxmox at this point because I could still run Windows as a VM while playing around and setting up a new drive pool. I’d like a setup that I can gradually upgrade because I don’t often have a full day to dedicate to these matters.
MergerFS still seems like a good fit for my media pool, simply only to solve an issue where one media type is filling a whole drive as another sits at 50% capacity. I’ve lost this data before and it was easy to recover by way of my preferred backup method (private torrent tracker with paid freeleech). A parity drive with SnapRaid might be a nice stop gap. I don’t think I feel confident enough with ZFS to potentially sacrifice uptime.
My dockers and server databases, however, are on a separate SSD that could benefit from ZFS. These files are backed up regularly so I can recover easily and I’d like as many failsafes as possible to protect myself. Having my Radarr database was indispensable when I lost a media drive a few weeks ago.
Proxmox is pretty much focused on ZFS, LXC containers and VMs. You want mergerFS and Docker. I say avoid Proxmox and go for Debian or another distro.
MergerFS and SnapRAID could be good for you. It’s not immediate parity like with ZFS RAID (You run a regular cronjob to calculate RAID parity) but it supports mismatched drive sizes, expansion of the pool at any time, and some other features that should be good for a media server where live parity isn’t critical.
Proxmox and TrueNAS are nice because they help manage ZFS and other remote management within a nice UI but really you can just use Debian with SSH and do the same stuff. DietPi has a few nice utilities on top of Debian (DDNS manager and CLI fstab utilities, for example)but not super necessary.
Personally I use TrueNAS but I also used DietPi/Debian for years and both have benefits and it really matters what your workflow is. OMV supports everything you want too (incouding SnapRAID) but takes extra setup which put me off.
Docker or LXC containers won’t hurt your performance btw. There’s supposedly some tiny overhead but both are designed to use the basic Linux system as much as possible: they’re way faster than on WSL. For hardware acceleration it’ll be deferred to the GPU for most things and there’s lots of documentation to set it up. The best thing about docker is that every application is kept separate to eachother - updates can be done incrementally and rollbacks are possible too!
Proxmox is Debian under the hood. It’s just a qemu and lxc management interface.
yeah, and qemu and lxc are very much legacy at this point. Stick with docker/podman/kubernetes for containers.
QEMU is legacy? Pray tell me how you’re running VMs on architectures other than x86 on modern computers without QEMU
Not QEMU in particular, poor phrasing on my part. I just mean setting up new environments that run applications on VMs.
I prefer some of my applications to be on VMs. For example, my observability stack (ELK + Grafana) which I like to keep separate from other environments. I suppose the argument could be made that I should spin up a separate k8s cluster if I want to do that but it’s faster to deploy directly on VMs, and there’s also less moving parts (I run two 50 node K8S clusters so I’m not averse to containers, just saying). Easier and relatively secure tool for the right job. Sure, I could mess with cgroups and play with kernel parameters and all of that jazz to secure k8s more but why bother when I can make my life easier by trusting Red Hat? Also I’m not yet running a k8s version that supports SELinux and I tend to keep it enabled.
Yeah I’m not saying everybody has to go and delete their infra, I just think that all new production environments should be k8s by default.
The production-scale Grafana LGTM stack only runs on Kubernetes fwiw. Docker and VMs are not supported. I’m a bit surprised that Kubernetes wouldn’t have enough availability to be able to co-locate your general workloads and your observability stack, but that’s totally fair to segment those workloads.
I’ve heard the argument that “kubernetes has more moving parts” a lot, and I think that is a misunderstanding. At a base level, all computers have infinite moving parts. QEMU has a lot of moving parts, containerd has a lot of moving parts. The reason why people use kubernetes is that all of those moving parts are automated and abstracted away to reduce the daily cognitive load for us operations folk. As an example, I don’t run manual updates for minor versions in my homelab. I have a k8s CronJob that runs renovate, which goes and updates my Deployments in git, and ArgoCD automatically deploys the changes. Technically that’s a lot of moving parts to use, but it saves me a lot of manual work and thinking, and turns my whole homelab into a sort of automated cloud service that I can go a month without thinking about.
I’m not sure if container break-out attacks are a reasonable concern for homelabs. See the relatively minor concern in the announcement I made as an Unraid employee last year when Leaky Vessels happened. Keep in mind that containerd uses cgroups under the hood.
Yeah, apparmor/selinux isn’t very popular in the k8s space. I think it’s easy enough to use them, plenty of documentation out there; but Openshift/okd is the only distribution that runs it out of the box.
By more moving parts I mean:
Running ElasticSearch on RHEL:
- add repo and dnf install elasticsearch.
- check SELinux
- write config
- firewall-cmd to open ports.
In k8s:
- grab elasticsearch container image
- edit variables in manifest (we use helm)
- depending on if the automatically configured SVC is good, leave it alone or edit it.
- write the VS and gateway (we use Istio)
- firewall-cmd to open ports
Maybe it’s just me but I find option 1 easier. Maybe I’m just lazy. That’s probably the overarching reason lol
You’re not using a reverse proxy on rhel, so you’ll need to also make sure that the ports you want are available, and set up a dns record for it, and set up certbot.
On k8s, I believe istio gateways are meant to be reused across services. You’re using a reverse proxy so the ports will already be open, so no need to use firewall-cmd. What would be wrong with the Service included in the elasticsearch chart?
It’s also worth looking at the day 2 implications.
For backups you’re looking at bespoke cronjobs to either rsync your database or clone your entire 100gb disk image, compared to either using velero or backing up your underlying storage.
For updates, you need to run system updates manually on rhel, likely requiring a full reboot of the node, while in kubernetes, renovate can handle rolling updates in the background with minimal downtime. Not to mention the process required to find a new repo when rhel 11 comes out.
right tool for the job mate, not everything works great in a container.
Also Proxmox is not legacy as its used a lot in homelabs and also some companys
I use proxmox to carve up my dedicated host with OVH, 3 of the vms run docker anyway.
I’m not saying it’s bad software, but the times of manually configuring VMs and LXC containers with a GUI or Ansible are gone.
All new build-outs are gitops and containerd-based containers now.
For the legacy VM appliances, Proxmox works well, but there’s also Openshift virtualization aka kubevirt if you want take advantage of the Kubernetes ecosystem.
If you need bare-metal, then usually that gets provisioned with something like packer/nixos-generators or cloud-init.
Sometimes, VMs are simply the better solution.
I run a semi-production DB cluster at work. We have 17 VMs running and it’s resilient (a different team handles VMWare and hardware)
I have 33 database servers in my homelab across 11 postgres clusters, all with automated barman backups to S3.
Here is the entire config for the db cluster that runs my Lemmy instance
This stuff is all automated these days.
Ah thanks, I’ll go through it!
Yes, but no. There is still a lot of places using old fashioned VMs, my company is still building VMs from an AWS ami and running ansible to install all the stuff we need. Some places will move to containers and that’s great, but containers won’t solve every problem
Yes, it’s fine to still have VMs, but you shouldn’t be building out new applications and new environments on VMs or LXC.
The only VMs I’ve seen in production at my customers recently are application test environments for applications that require kernel access. Those test environments are managed by software running in containers, and often even use something like Openshift Virtualization so that the entire VM runs inside a container.
but you shouldn’t be building out new applications and new environments on VMs or LXC
That’s a bold statement, VMs might be just fine for some.
Use what ever is best for you, if thats containers great. If that’s a VM, sure. Just make sure you keep it secure.
Some of us don’t build applications, we use them as built by other companies. If we’re really unlucky they refuse to support running on a VM.
Yeah, that’s fair. I have set up Openshift Virtualization for customers using 3rd party appliances. I’ve even worked on some projects where a 3rd party appliance is part of the original spec for the cluster, so installing Openshift Virtualization to run VMs is part of the day 1 installation of the Kubernetes cluster.
Why would you install a GUI on a VM designated to run a Docker instance?
You should take a serious look at what actual companies run. It’s typically nested VMs running k8s or similar. I run three nodes, with several VMs (each running Docker, or other services that require a VM) that I can migrate between nodes depending on my needs.
For example: One of my nodes needed a fan replaced. I migrated the VM and LXC containers it hosted to another node, then pulled it from the cluster to do the job. The service saw minimal downtime, kids/wife didn’t complain at all, and I could test it to make sure it was functioning properly before reinstalling it into the cluster and migrating things back at a more convenient time.
I’m a DevOps/ Platform Engineering consultant, so I’ve worked with about a dozen different customers on all different sorts of environments.
I have seen some of my customers use nested VMs, but that was because they were still using VMware or similar for all of their compute. My coworkers say they’re working on shutting down their VMware environments now.
Otherwise, most of my customers are running Kubernetes directly on bare metal or directly on cloud instances. Typically the distributions they’re using are Openshift, AKS, or EKS.
My homelab is all bare metal. If a node goes down, all the containers get restarted on a different node.
My homelab is fully gitops, you can see all of my kubernetes manifests and nixos configs here:
You are going to what, install Kubernetes on every node?
It is far easier and more flexible to use VMs and maybe some VM templates and Ansible.
Yes.
It is not easier to use Ansible. My customers are trying to get rid of Ansible.
Agreed.
I run podman w/ rootless containers, and it works pretty well. Podman is extra nice in that it has decent suppor for kubernetes, so there’s a smooth transition path from podman -> kubernetes if you ever want/need it. Docker works well too, and
docker composeis pretty simple to get into.Yeah, Kubernetes is more automated and expandable, but docker compose has a ton of good examples and it’s really easy to get into as a beginner.
Kubernetes is also designed for clustered workloads, so if you are mostly hosting on one or two machines, YAGNI applies.
I recommend people start w/ docker compose due to documentation, but I personally am switching to podman quadlets w/ rootless containers.
Yeah, definitely true.
I’m a big fan of single-node kubernetes though, tbh. Kubernetes is an automation platform first and foremost, so it’s super helpful to use Kubernetes in a homelab even if you only have one node.
What’s so nice about it? Have you tried quadlets or docker compose? Could you give a quick comparison to show what you one like about it?
Sure!
I haven’t used quadlets yet, but I did set up a few systemd services for containers back in the day before quadlets came out. I also used to use docker compose back in 2017/2018.
Docker compose and Kubernetes are very similar as a homelab admin. Docker compose syntax is a little less verbose, and it has some shortcuts for storage and networking. But that also means it’s less flexible if you are doing more complex things. Docker compose doesn’t start containers on boot by default I think(?) which is pretty bad for application hosting. Docker-compose has no way of automatically deploying from git like ArgoCD does.
Kubernetes also has a lot of self-healing automation, like health checks that can either disable the load balancer and/or restart the container if an app is failing, automatic killing of containers when resources are low, preventing the scheduling of new containers when resources are low, gradual roll-out of containers so that the old version of a container doesn’t get killed until the new version is up and healthy (helpful in case the new config is broken), mounting secrets as files in a container, and automatic retry on failed containers.
There’s also a lot of ubiquitous automation tools in the Kubernetes space, like cert-manager for setting up certificates (both ACME and local CA), Ingress for setting up reverse proxy, CNPG for setting up postgres clusters with automated backups, and first-class instrumentation/integration with prometheus and loki (both were designed for kubernetes first).
The main downsides with Kubernetes in a homelab is that there is about a 1-2GiB RAM overhead for small clusters, and most documentation and examples are written for docker-compose, so you have to convert apps into a Deployment (you get used to writing deployments for new apps though). I would say installing things like Ingress or CNPG is probably easier than installing similar reverse-proxy automations on Docker-compose, though.
What are you going to run containers on? You need VMs to power everything.
I dont have any VMs running in my homelab.
Most of my customers run their Kubernetes nodes either on bare metal, or on a cloud provisioned VM from AWS/GCP/Azure etc
None of your listed use cases will even come close to taxing the 6600k. It’s going to probably sit happily in idle states most of the time.
Proxmox also has great snapshotting and backup features. Makes it easier to mess around with your containers/VMs without worrying too much.
Only when using zfs, which op is not.
I don’t know about your first need (“MergerFS”) but if you find useful, I have an old Intel NUC 6i3SYH (i3-6100U) with 16Gb RAM and I was running with Windows 10 for Plex+Arr and also HomeAssistant in VirtualBox. I was running into issues until I switched to Proxmox. Now I’m running Proxmox to run Docker with a bunch of containers (plex+arr and others) and also a virtual machine which has HomeAssistant and everything was smooth. I have to say that there is a learning curve, but it’s very stable.
Seconding this, I’m currently running Proxmox on 3 small NUC-type PCs (two Dell Optiplexes and a Topton from AliExpress). The Topton has a slower Celeron, the two Dells have a i5-6500 and i3-8100t and are both very snappy running a few different containers and VMs (including HomeAssistant).
I use OpenMediaVault to run something similar. It’s a headless Debian distribution with web based config. Takes a bit of work but I like it.
My server runs Debian VMs in Proxmox on an i7-2600 which has a lower benchmark than the 6600k. I also used the Perfect Media Server guide, and have 2 x 8T data drives pooled with MergerFS with 1 for snapraid parity, these are passed through to the main VM from Proxmox using ‘qm set’. One thing I would often forget after deleting/restoring this VM was to run qm set again after restore, ensuring it has the flag to not back up those drives or else backups will fail and I have to go uncheck the backup option on each drive to fix it.
If I need to spin up another VM for tinkering it’s easy enough to mount the NFS share as a volume with docker compose. Proxmox rarely shows CPU usage go above 50% (average is 10%) and this handles the whole *arr stack plus usenet and torrents in a single VM and compose file. I don’t have GPU passthrough set up because the motherboard on this older rig didn’t support IOMMU, never had issues with Plex or Jellyfin transcoding for Chromecast. I might build a new rig with GPU passthrough support to buffer media faster and selfhost LLMs when I get around to it.
Your CPU should be perfectly capable of that. I ran Proxmox with some VMs and containers on an i5-2400 with 16GB RAM just fine.
You could run on bare Debian as well but virtualization will give you more flexibility. If you get a Zigbee Dongle or the like, you can pass it through to the VM Home Assistant is running in.
I don’t know MergeFS but usually the recommendation is ZFS.
would agree the hardware would run everything fine
Not calling you out specifically OP, but can someone tell me why this is a thing on the internet?
multiple 12GB drives
GB??? I assume TB automatically when people say this but it still is a speedbreaker when I’m thinking about the post.
Good catch, yes my drives are 12TB. My brain is still stuck in 2005. :)
Promox runs on debian. But anyway you will be surprised about proxmox can run in limited hardware. I have it running in a garbage mini PC and an old notebook :D
Proxmox is Debian. :-)
I do always suggest installing Debian first, and then installing Proxmox on top. This allows you to properly set up your disks, and networking as needed, as the Proxmox installer is a bit limited: https://pve.proxmox.com/wiki/Install_Proxmox_VE_on_Debian_12_Bookworm
Once you have it up and running, have a look at the CT Templates. There’s a whole set of pre-configured templates from TurnkeyLinux (again, debian+) that make it trivial to set up all kinds of services in lightweight LXC Containers.
For Home Assistant a VM is your best bet, as it makes setting up connectivity way easier than messing with docker networking. It also allows easy USB passthrough, for things like ZWave/Zigbee/Bluetooth adapters.I would just install Proxmox since it is way easier
Also last time I checked the Debian installer didn’t support ZFS
I do always suggest installing Debian first, and then installing Proxmox on top.
Correct me if I’m wrong, but isn’t Proxmox it’s own OS unto itself? What would be the advantage of installing Proxmox ‘on top of’ Debian when it’s Debian already as you pointed out?
You have some options that aren’t in the installer e.g. full disk encryption
hmmmm. Wouldn’t you have to remove the Debian kernal and use the Proxmox kernal? Sorry, not trying to be obtuse, I just have never installed Proxmox ‘on top’ of Debian. I always opted for the clean install.
Yes, but that’s a supported way to install Proxmox.
https://pve.proxmox.com/wiki/Install_Proxmox_VE_on_Debian_12_Bookworm
Awww man. I hope you didn’t think I was questioning you. I was just curious and I never knew that or would have guessed. I learned something. Thanks.
Don’t worry, I didn’t :)
It’s also, I find, much more widely supported on a wider variety of hardware and with easier config automation.
OP, I’m running Proxmox on and old Dell T320 /32gb RAM. I am not having any real issues doing so. I run Docker and a handful of Docker containers. I’m really not into the arr stack, but I wouldn’t think you’d have much issue.
