Mastodawn

Very excited to have gotten through a pretty big lift + refactor in the #homelab over the last few weeks. For a while, I’ve had one foot in my old, disorganized, scattered Docker services, and one foot in a new, fully automated, #gitops pseudo-prod for home.

Well, no more. I’ve cleaned up all of the old shit. Centralized config, linted, backed up, refactored, and smoke tested the core platform. It completely spins up and down, with one Playbook, data fully backed up, etc.

Hardware:
- 3x #amd mini PC’s w/32GB RAM
- #synology DS1813+
- 1x Intel miniPC for #homeassistant OS
- TP-Link #Omada gateway, PoE switch, and AP’s

My core stack is, going from hardware to workload lifecycle and access:
- #proxmox
- #fedora #fedora_coreos VM’s
- #k3s #Rancher #kubernetes
- #nfs subdir provisioner
- #ansible
- #argocd
- #Tailscale
- #1password Kubernetes Operator
- #velero

I just did a DR drill with my kids’ #minecraft servers, and it worked great, and I’m super pumped with #selfhosting #paperlessngx!

Those are the first family-facing apps that I’ve deployed, and I’ve got a laundry list of more services on the way.

It was far, far more work than it probably needed to be…but I also can fully trust it with our most important personal docs, which was priority one.

Thib Nov 12

Pretty unimpressed that Velero encrypts volumes backups, but stores Secret resources in plain text on S3.

Am I missing a way to encrypt them?

#homelab #kubernetes #backups #velero

Saustrup Oct 16

Having recently experienced a rather horrible #Kubernetes crash, I'm looking for #backup solutions. We're good with PostgreSQL since we're using #CNPG with remote transaction logs to an offsite #S3 bucket. I need something for volumes and maybe Kubernetes resources. #Longhorn offers S3 backups for it's own volumes, but for other #CSI like local #OpenEBS, maybe #Velero? Thoughts?

https://velero.io/

Velero

Dqpcoxeas Sep 27

#darktable #photography #landscape #fotografie #foto #clouds #sky #cielo #Vigo #galicia #velero #navegacion #mar #sea #ocean #sailboat #sailing

Rachel Sep 12, 2025

Homelab project idea: Standby disaster recovery/Continuity of Operations node

The parent's fileserver (PFS) will be running Proxmox, and will have plenty of headroom. The idea is to run a simple single-node cluster in a VM (likely Talos) and configure Velero to do a restore of critical applications to this single-node cluster.

I'm not sure how to handle storage classes. The DR node won't have Ceph, I'm thinking maybe configuring a local storage provisioner and making fake storage classes (i.e. a "cephfs" storage class that lies and points to a local disk).

Cilium/CNI would be configured in a much more basic manner as well maybe I could also do the fake thing and map the "cilium" ingress to something else, or just configure it in L2 mode. All I would need to do is switch some DNS records and I'd have only lost changes since the last backup.

The first victory would be a proof of concept but it might be possible to have a daily task that overwrites the DR cluster with the most recent successful backup.

Since the home automation and media streaming/acquisition apps won't be set to run on the DR node, a so goes reasonably sized VM should be able to handle things with no problem.

This would also put me at ease that if needed I could move my self-hosted platform to an event smaller significantly more portable/lower power physical machine, maybe even a single odroid H4 with 2-4 disks

#Kubernetes #Homelab #Velero #Backups #Selfhosted

Madic Sep 2, 2025

I was remembered, that behind velero backup Broadcom is. Time to ditch velero before Broadcom does Broadcom things and test Volsync for backups.

#kubernetes #backup #velero #volsync

Rachel Jul 14, 2025

The Velero backups are working except for the CephFS volumes

I have no idea why it just throws timeout errors but I can create a manual snapshot which doesn't take that long?

I'm so close to getting past this step but I'm not exactly where to look next
#Kubernetes #Velero #Ceph #Backups