New blog post! I'm preparing some topology changes in my home lab, so I had to spin up a new Talos Linux control plane node on my Proxmox Cluster. As it already happened before, enabling secure boot required some manual setup. This blog post might help if want to do the same.
#HomeLab #TalosLinux #Proxmox #Blog @homelab
https://mteixeira.wordpress.com/2026/02/01/booting-talos-on-a-proxmox-vm/
#TalosLinux
New in #TalosLinux & Omni
→ Multi-doc configs for networking
→ Talos Linux cluster imports for Omni
→ Pre-seeded images
→ Built-in OOM handling
Shells are for snails 🐌 not servers #TalosLinux
Alright, using #talosLinux with #terraform and #libvirt to build out all the #kubernetes infra to run the #selfHosted bits.
So I have 2 5950x with 64gb of ram and 4tb of storage. Then a 3950x with 1tb and 32g, and finally some small older boxes.
My desktop w the 5950 used to be a gaming machine but the steam deck has 100% replaced that functionality.
Eventually I may just run talos natively on either of the larger boxes but libvirt is a much nicer api.
RE: https://hachyderm.io/@badnetmask/115698872870594708
Home lab architecture update: two of the Lenovo boxes in the picture below became Talos nodes, together with the Framework Desktop. The third Lenovo became a Proxmox node.
Yep, I'm back on the Proxmox bandwagon.
Now I have a bit of everything: Talos running Kubernetes, Proxmox running VM/LXC and Raspberry Pi running Docker.
At some point the three Raspberry Pis are going away, and the services are going to be distributed between Talos and Proxmox, then they are going to become back-end compute for shenanigans (like a Home Assistant dashboard, for example).
RE: https://hachyderm.io/@siderolabs/115853943771571011
When someone asks me "why Talos and not K3s", I usually simplify and say "I don't want an extra OS layer". Well. This blog post goes beyond my simplistic explanation. I hope it helps.
Poll time! I think I overdid it: I have 7 storage classes in my cluster. There are 3 for local storage (Longhorn) and 4 for network storage (Synology). The good thing is that it gives me flexibility.
Storage classes:
- longhorn (reclaim=delete) - replicated across 2 nodes
- longhorn-retain - also replicated
- longhorn-fw1-local (reclaim=delete) - not replicated, exists only in a dedicated NVMe in the fw1 node, with much more space than the replicated classes
- {iscsi,nfs}-{delete,retain} - self-explanatory
I'm wondering if the classes with reclaim policy retain is overkill, since any important data that I care about is backed up outside the cluster anyway. Feel free to cast a vote below, but I would appreciate if you also comment why.
New blog post! How to configure the Lenovo M720Q Tiny (and probably any other Lenovo) to boot Talos Linux in Secure Boot mode.
#HomeLab #TalosLinux #Lenovo #blog @homelab
https://mteixeira.wordpress.com/2025/12/26/booting-talos-linux-on-a-lenovo-m720q-tiny/
One interesting side-effect of a cluster rebuild is finding stuff that was done manually and not added to GitOps. :blobfoxlaughsweat:
#HomeLab #TalosLinux @homelab
The Talos upgrade to v1.12.0 GA (and Kubernetes 1.35.0 GA) was easy and flawless. Also a good test to see if things re-schedule properly (which they did).
@homelab
After a lot of sweat and tears, I have a fully functional 3-node cluster!
I took this opportunity to make a bunch of changes in the configs, so I just nuked everything and started over. That whole experience taught me a lot of new things, and helped surface a lot of mistakes I made during my "organic" initial setup.
Most of the learnings are going to surface here as Mastodon posts. At least one will be a proper blog post. Unsure how many others can be.
You will notice on the screenshot below that I am running Talos 1.12.0 rc1. That's because I want the newest kernel to test some improvements with the integrated GPU. In theory the GA version will come tomorrow, so it should be a very simple upgrade.
The holiday break starts now! Time to upgrade the home lab adding new nodes to my Talos cluster.
(due to a snafu, this is a re-post)
New blog post! After reading @mmeier's blog post about monitoring his Kubernetes cluster certificates, I decided to take a look at how this is done with Talos, and learned a lot from it. You can read my solution in my blog post, as well as Michael's (which I link in my post, and also below in this thread.
#HomeLab #TalosLinux #StepCA #Certificates #Kubernetes #Blog @homelab
https://mteixeira.wordpress.com/2025/12/07/monitoring-the-kubernetes-certificates-on-a-talos-cluster/
(To the anti-AI people: please feel free to ignore/skip this post.)
New blog post! A few people have asked me how I run AI/LLM tools (ollama and llama.cpp) in my Talos Kubernetes cluster, so I decided to give it a proper explanation. This is quite a long and complex post, but I hope it's useful for someone.
#HomeLab #TalosLinux #Framework #ollama #llamacpp #AI #LLM #AMD #StrixHalo @homelab
"Kubernetes is overkill for a Home Lab", exhibit 512 (made up number): KubeVirt tagged my nodes with 200+ tags, and that caused a metrics overload, so I had to increase the pod requests in the collectors, which increased my power usage.
Then a few days later I started getting alerts of expired certs, and realized that KubeVirt has been creating thousands of one-off certificates, but not reusing them. Increasing the certificate duration not only stopped the alerts, but also *reduced the power usage*.
All this trouble and I don't even have a single VM running yet. I hope this is all worth the trouble.
#HomeLab #TalosLinux #KubeVirt @homelab
https://github.com/kubevirt/kubevirt/issues/13548
The spice must flow, and the home lab must grow! I have acquired 3 refurbished Lenovo M720q Tiny.
In the picture below you will see my Framework Desktop on the right, which is currently running my single-node Talos Kubernetes cluster. Two of the Lenovo will be added as nodes to the cluster. I still haven't decided what I am going to do with the third, but most likely it will become a NAS at some point.
The Lenovos have been acquired via a store named "Untrusted Source". Don't let the name fool you. They're incredibly friendly and knowledgeable about the craft of refurbishing hardware (see link below).
A 3 node 10g capable SFF #kubernetes cluster that idles around 25-35w. Not too shabby. #TalosLinux #homelab
And now I have KubeVirt installed in my Talos cluster, with a test VM provisioned. Need to finish up some tweaks, then provision a real VM for real use. That's probably going to become a blog post (at some point). 😄
Edit: I realized I need to extend one extra VLAN into my cluster for one of the VMs. Oh boy, that's gonna be a hoot!
"Kubernetes is overkill for a Home Lab", exhibit 493 (made up number): I'm trying to make Pi-Hole egress to the internet via the loadbalancer IP, rather than the node IP (default). To do that, I need to implement Cilium Egress Gateway, which is horribly complicated, and will be used for a one-off. But if I don't do so, I need to review all my firewall rules, which only allow very few IP addresses in my network to query external DNS servers. Ugh. I'm starting to reconsider my sanity. What am I getting myself into? :picardfacepalm:
Edit: nothing like a good night of sleep. I found a solution using traffic policies in my router, so I no longer need the complexity of the k8s egress gateway. 👍
So, apparently there is a bug in Kubernetes 1.34.1, which prevents some metrics to be properly registered. Some of those are important for storage/PVC monitoring, so they were reestablished in 1.34.2.
Upgrading my cluster might allow me to (at least partially) fix my PVC monitoring blind spot. Now I just need to find some free time in my life to do that upgrade and verify this actually solves my problem.
The funny part is that this all appears to be related to DRA (Dynamic Resource Allocation), which was introduced in 1.34.0, and extensively mentioned in various talks at KubeCon, so I was very interested in deep diving into it. What a coincidence.
#HomeLab #TalosLinux #ZFS #Kubernetes #KubeCon @homelab
https://github.com/kubernetes/kubernetes/issues/133847
Client Info
Server: https://mastodon.social
Version: 2025.07
Repository: https://github.com/cyevgeniy/lmst