+1 -2

README.md

reviewed

··· 8 8 9 9 - Drimor: an ASUS NUC 14 running Debian with several KVM guests: 10 10 - Ayzinith: Alpine VM running [pi-hole](https://github.com/pi-hole/pi-hole) 11 11 - - Yix: Ubuntu VM running [MinIO](https://github.com/minio/minio) for S3 storage 12 12 - - Persepolis: Home Assistant OS VM 11 11 + - ~~Persepolis: Home Assistant OS VM~~ (currently broken after an SSD failure) 13 12 - Proteus Quartus: A [Talos](https://talos.dev) node 14 13 - Proteus Secundus: an ASUS NUC 15 running a [Talos](https://talos.dev) node on bare metal 15 14 - Paleon: A [Scaleway](https://scaleway.com/) VPS running an NGINX reverse proxy that exposes certain services over Tailscale to the outside world.

+3 -2

kubernetes/apps/office/immich/helmrelease.yaml

reviewed

··· 21 21 capabilities: 22 22 drop: [ ALL ] 23 23 pod: 24 24 - nodeSelector: 25 25 - intel.feature.node.kubernetes.io/gpu: "true" 26 24 securityContext: 27 25 runAsUser: 1000 28 26 runAsGroup: 1000 ··· 99 97 machine-learning: 100 98 controllers: 101 99 main: 100 100 + pod: 101 101 + nodeSelector: 102 102 + intel.feature.node.kubernetes.io/gpu: "true" 102 103 containers: 103 104 main: 104 105 image:

+1 -1

procedures/disaster-recovery/cnpg.md

reviewed

··· 8 8 9 9 ## Semi-broken 10 10 11 11 - These are the easy ones: run `kubectl cnpg destroy -n postgres <name of cluster> <generation of broken cluster>! 11 11 + These are the easy ones: run `kubectl cnpg destroy -n postgres <name of cluster> <generation of broken cluster>`! 12 12 You'll note the failed volumes disappear from the Longhorn UI as the broken instance is deleted by the CNPG controller. 13 13 The new instance created afterwards will immediately sync & you'll have two copies of your data again. 14 14