Compare commits

..

18 Commits

Author SHA1 Message Date
Gitea Actions Bot
629aa1e50f Auto-update README with current k8s applications
All checks were successful
Terraform / Terraform (pull_request) Successful in 20s
Generated by CI/CD workflow on 2026-03-07 21:27:33

This PR updates the README.md file with the current list of applications found in the k8s/ directory structure.
2026-03-07 21:27:33 +00:00
0704ac33c7 Enable GPU time-slicing for ComfyUI and Ollama
All checks were successful
Update Kubernetes Services Wiki / Generate and Update K8s Wiki (push) Successful in 8s
Check with kubeconform / lint (push) Successful in 6s
Auto-update README / Generate README and Create MR (push) Successful in 7s
2026-03-07 21:27:10 +00:00
a499210fa2 Change ArgoCD project to apps
All checks were successful
Update Kubernetes Services Wiki / Generate and Update K8s Wiki (push) Successful in 7s
Check with kubeconform / lint (push) Successful in 6s
Auto-update README / Generate README and Create MR (push) Successful in 6s
2026-03-07 21:24:51 +00:00
ca15b9a569 Add toleration for desktop node to ComfyUI deployment
Some checks failed
Check with kubeconform / lint (push) Has been cancelled
Update Kubernetes Services Wiki / Generate and Update K8s Wiki (push) Has been cancelled
Auto-update README / Generate README and Create MR (push) Has been cancelled
2026-03-07 21:24:39 +00:00
abae3690f9 Add ComfyUI manifests and ArgoCD application
All checks were successful
Update Kubernetes Services Wiki / Generate and Update K8s Wiki (push) Successful in 8s
Check with kubeconform / lint (push) Successful in 7s
Auto-update README / Generate README and Create MR (push) Successful in 6s
2026-03-07 21:24:02 +00:00
a3540b1706 Fixed n8n worker
All checks were successful
Update Kubernetes Services Wiki / Generate and Update K8s Wiki (push) Successful in 9s
Check with kubeconform / lint (push) Successful in 7s
Auto-update README / Generate README and Create MR (push) Successful in 6s
2026-03-07 10:34:54 +00:00
f11c561f3a fix(ollama): update to standard kustomize patches format and target ollama namespace
All checks were successful
Update Kubernetes Services Wiki / Generate and Update K8s Wiki (push) Successful in 9s
Check with kubeconform / lint (push) Successful in 6s
Auto-update README / Generate README and Create MR (push) Successful in 6s
2026-03-07 01:03:25 +00:00
0c2ff6cc3b fix(ollama): add runtimeClassName nvidia to enable GPU usage
All checks were successful
Update Kubernetes Services Wiki / Generate and Update K8s Wiki (push) Successful in 9s
Check with kubeconform / lint (push) Successful in 7s
Auto-update README / Generate README and Create MR (push) Successful in 6s
2026-03-07 01:02:18 +00:00
79d991da1e feat(ollama): add local persistent volume for model storage on uk-desktop
All checks were successful
Update Kubernetes Services Wiki / Generate and Update K8s Wiki (push) Successful in 8s
Check with kubeconform / lint (push) Successful in 6s
Auto-update README / Generate README and Create MR (push) Successful in 6s
2026-03-07 00:57:13 +00:00
49b7e0a8c2 feat(openwebui): schedule sub-components on master node
All checks were successful
Update Kubernetes Services Wiki / Generate and Update K8s Wiki (push) Successful in 8s
Check with kubeconform / lint (push) Successful in 6s
Auto-update README / Generate README and Create MR (push) Successful in 8s
2026-03-07 00:55:56 +00:00
67b094ae70 fix(openwebui): set clusterDomain to cluster.local to fix redis url
All checks were successful
Update Kubernetes Services Wiki / Generate and Update K8s Wiki (push) Successful in 7s
Check with kubeconform / lint (push) Successful in 6s
Auto-update README / Generate README and Create MR (push) Successful in 7s
2026-03-07 00:49:24 +00:00
c9d29d91d2 Updated openwebUI vars
All checks were successful
Update Kubernetes Services Wiki / Generate and Update K8s Wiki (push) Successful in 8s
Check with kubeconform / lint (push) Successful in 6s
Auto-update README / Generate README and Create MR (push) Successful in 6s
2026-03-07 00:41:54 +00:00
ab
8b12988aaf Update k8s/apps/ollama/kustomization.yaml
Some checks failed
Update Kubernetes Services Wiki / Generate and Update K8s Wiki (push) Successful in 7s
Check with kubeconform / lint (push) Successful in 6s
Auto-update README / Generate README and Create MR (push) Failing after 7m22s
2026-03-07 00:27:11 +00:00
f9943203fc fix: Add workload=desktop toleration to ollama
All checks were successful
Update Kubernetes Services Wiki / Generate and Update K8s Wiki (push) Successful in 7s
Check with kubeconform / lint (push) Successful in 5s
Auto-update README / Generate README and Create MR (push) Successful in 7s
2026-03-06 23:33:55 +00:00
bbf2ea3911 fix: Add toleration for workload=desktop:NoSchedule taint on uk-desktop
All checks were successful
Update Kubernetes Services Wiki / Generate and Update K8s Wiki (push) Successful in 7s
Check with kubeconform / lint (push) Successful in 5s
Auto-update README / Generate README and Create MR (push) Successful in 6s
2026-03-06 23:31:23 +00:00
f62a196d1f fix: Remove invalid config.map+config.name in nvidia-device-plugin values
All checks were successful
Update Kubernetes Services Wiki / Generate and Update K8s Wiki (push) Successful in 8s
Check with kubeconform / lint (push) Successful in 7s
Auto-update README / Generate README and Create MR (push) Successful in 6s
2026-03-06 23:29:56 +00:00
23e5148814 feat: Add NVIDIA GPU config and update Ollama to use it
All checks were successful
Update Kubernetes Services Wiki / Generate and Update K8s Wiki (push) Successful in 7s
Check with kubeconform / lint (push) Successful in 7s
Auto-update README / Generate README and Create MR (push) Successful in 7s
2026-03-06 23:24:40 +00:00
e1bd487ec9 Added N8N_PROXY_HOPS to n8n
All checks were successful
Update Kubernetes Services Wiki / Generate and Update K8s Wiki (push) Successful in 7s
Check with kubeconform / lint (push) Successful in 16s
Auto-update README / Generate README and Create MR (push) Successful in 6s
2026-03-06 15:56:37 +00:00
20 changed files with 354 additions and 11 deletions

View File

@@ -16,6 +16,7 @@ ArgoCD homelab project
| **authentik** | [![authentik](https://ag.hexor.cy/api/badge?name=authentik&revision=true)](https://ag.hexor.cy/applications/argocd/authentik) |
| **cert-manager** | [![cert-manager](https://ag.hexor.cy/api/badge?name=cert-manager&revision=true)](https://ag.hexor.cy/applications/argocd/cert-manager) |
| **external-secrets** | [![external-secrets](https://ag.hexor.cy/api/badge?name=external-secrets&revision=true)](https://ag.hexor.cy/applications/argocd/external-secrets) |
| **gpu** | [![gpu](https://ag.hexor.cy/api/badge?name=gpu&revision=true)](https://ag.hexor.cy/applications/argocd/gpu) |
| **kube-system-custom** | [![kube-system-custom](https://ag.hexor.cy/api/badge?name=kube-system-custom&revision=true)](https://ag.hexor.cy/applications/argocd/kube-system-custom) |
| **kubernetes-dashboard** | [![kubernetes-dashboard](https://ag.hexor.cy/api/badge?name=kubernetes-dashboard&revision=true)](https://ag.hexor.cy/applications/argocd/kubernetes-dashboard) |
| **longhorn** | [![longhorn](https://ag.hexor.cy/api/badge?name=longhorn&revision=true)](https://ag.hexor.cy/applications/argocd/longhorn) |
@@ -37,6 +38,7 @@ ArgoCD homelab project
| Application | Status |
| :--- | :---: |
| **comfyui** | [![comfyui](https://ag.hexor.cy/api/badge?name=comfyui&revision=true)](https://ag.hexor.cy/applications/argocd/comfyui) |
| **gitea** | [![gitea](https://ag.hexor.cy/api/badge?name=gitea&revision=true)](https://ag.hexor.cy/applications/argocd/gitea) |
| **greece-notifier** | [![greece-notifier](https://ag.hexor.cy/api/badge?name=greece-notifier&revision=true)](https://ag.hexor.cy/applications/argocd/greece-notifier) |
| **hexound** | [![hexound](https://ag.hexor.cy/api/badge?name=hexound&revision=true)](https://ag.hexor.cy/applications/argocd/hexound) |

20
k8s/apps/comfyui/app.yaml Normal file
View File

@@ -0,0 +1,20 @@
apiVersion: argoproj.io/v1alpha1
kind: Application
metadata:
name: comfyui
namespace: argocd
spec:
project: apps
destination:
namespace: comfyui
server: https://kubernetes.default.svc
source:
repoURL: ssh://git@gt.hexor.cy:30022/ab/homelab.git
targetRevision: HEAD
path: k8s/apps/comfyui
syncPolicy:
automated:
selfHeal: true
prune: true
syncOptions:
- CreateNamespace=true

View File

@@ -0,0 +1,50 @@
apiVersion: apps/v1
kind: Deployment
metadata:
name: comfyui
namespace: comfyui
labels:
app: comfyui
spec:
replicas: 1
selector:
matchLabels:
app: comfyui
template:
metadata:
labels:
app: comfyui
spec:
runtimeClassName: nvidia
tolerations:
- key: workload
operator: Equal
value: desktop
effect: NoSchedule
nodeSelector:
kubernetes.io/hostname: uk-desktop.tail2fe2d.ts.net
# Fix permissions mismatch usually happening when mapping host paths
securityContext:
runAsUser: 0
containers:
- name: comfyui
image: ghcr.io/ai-dock/comfyui:latest-cuda
imagePullPolicy: IfNotPresent
env:
- name: CLI_ARGS
value: "--listen 0.0.0.0 --port 8188"
ports:
- containerPort: 8188
name: http
protocol: TCP
resources:
limits:
nvidia.com/gpu: 1
volumeMounts:
- name: data
# For ai-dock images, /workspace is the persistent user directory
mountPath: /workspace
volumes:
- name: data
persistentVolumeClaim:
claimName: comfyui-data-pvc

View File

@@ -0,0 +1,9 @@
apiVersion: kustomize.config.k8s.io/v1beta1
kind: Kustomization
resources:
- namespace.yaml
- local-pv.yaml
- pvc.yaml
- deployment.yaml
- service.yaml

View File

@@ -0,0 +1,22 @@
apiVersion: v1
kind: PersistentVolume
metadata:
name: comfyui-data-pv
spec:
capacity:
storage: 200Gi
volumeMode: Filesystem
accessModes:
- ReadWriteOnce
persistentVolumeReclaimPolicy: Retain
storageClassName: local-path
local:
path: /data/comfyui
nodeAffinity:
required:
nodeSelectorTerms:
- matchExpressions:
- key: kubernetes.io/hostname
operator: In
values:
- uk-desktop.tail2fe2d.ts.net

View File

@@ -0,0 +1,4 @@
apiVersion: v1
kind: Namespace
metadata:
name: comfyui

12
k8s/apps/comfyui/pvc.yaml Normal file
View File

@@ -0,0 +1,12 @@
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
name: comfyui-data-pvc
namespace: comfyui
spec:
accessModes:
- ReadWriteOnce
storageClassName: local-path
resources:
requests:
storage: 200Gi

View File

@@ -0,0 +1,15 @@
apiVersion: v1
kind: Service
metadata:
name: comfyui
namespace: comfyui
labels:
app: comfyui
spec:
ports:
- name: http
port: 8188
targetPort: 8188
protocol: TCP
selector:
app: comfyui

View File

@@ -85,10 +85,14 @@ spec:
value: "queue"
- name: QUEUE_BULL_REDIS_HOST
value: "n8n-redis"
- name: QUEUE_BULL_REDIS_PORT
value: "6379"
- name: NODE_ENV
value: "production"
- name: WEBHOOK_URL
value: "https://n8n.hexor.cy/"
- name: N8N_PROXY_HOPS
value: "1"
- name: GENERIC_TIMEZONE
value: "Europe/Moscow"
- name: TZ
@@ -130,7 +134,7 @@ spec:
memory: 512Mi
limits:
cpu: 4000m
memory: 2048Gi
memory: 2048Mi
livenessProbe:
httpGet:
path: /healthz

View File

@@ -0,0 +1,84 @@
---
apiVersion: apps/v1
kind: Deployment
metadata:
name: n8n-worker
labels:
app: n8n
component: worker
spec:
replicas: 2
selector:
matchLabels:
app: n8n
component: worker
template:
metadata:
labels:
app: n8n
component: worker
spec:
serviceAccountName: n8n
containers:
- name: n8n-worker
image: n8nio/n8n:latest
command:
- n8n
- worker
env:
- name: HOME
value: "/home/node"
- name: N8N_ENFORCE_SETTINGS_FILE_PERMISSIONS
value: "true"
- name: EXECUTIONS_MODE
value: "queue"
- name: QUEUE_BULL_REDIS_HOST
value: "n8n-redis"
- name: QUEUE_BULL_REDIS_PORT
value: "6379"
- name: NODE_ENV
value: "production"
- name: GENERIC_TIMEZONE
value: "Europe/Moscow"
- name: TZ
value: "Europe/Moscow"
- name: DB_TYPE
value: "postgresdb"
- name: DB_POSTGRESDB_HOST
value: "psql.psql.svc"
- name: DB_POSTGRESDB_DATABASE
value: "n8n"
- name: DB_POSTGRESDB_USER
valueFrom:
secretKeyRef:
name: credentials
key: username
- name: DB_POSTGRESDB_PASSWORD
valueFrom:
secretKeyRef:
name: credentials
key: password
- name: N8N_ENCRYPTION_KEY
valueFrom:
secretKeyRef:
name: credentials
key: encryptionkey
volumeMounts:
- name: n8n-data
mountPath: /home/node/.n8n
resources:
requests:
cpu: 500m
memory: 512Mi
limits:
cpu: 2000m
memory: 2048Mi
volumes:
- name: n8n-data
persistentVolumeClaim:
claimName: n8n-data
securityContext:
runAsUser: 1000
runAsGroup: 1000
runAsNonRoot: true
fsGroup: 1000

View File

@@ -10,7 +10,7 @@ resources:
- paddleocr-deployment.yaml
- paddleocr-service.yaml
- deployment-main.yaml
# - deployment-worker.yaml
- deployment-worker.yaml
- deployment-runner.yaml
- service.yaml
- ingress.yaml

View File

@@ -3,19 +3,23 @@ kind: Kustomization
resources:
- external-secrets.yaml
- local-pv.yaml
helmCharts:
- name: ollama
repo: https://otwld.github.io/ollama-helm/
version: 0.4.0
version: 1.49.0
releaseName: ollama
namespace: ollama
valuesFile: ollama-values.yaml
includeCRDs: true
- name: open-webui
repo: https://helm.openwebui.com/
version: 8.14.0
version: 12.8.1
releaseName: openweb-ui
namespace: ollama
valuesFile: openweb-ui-values.yaml
includeCRDs: true
includeCRDs: true
patches:
- path: patch-runtimeclass.yaml

View File

@@ -0,0 +1,22 @@
apiVersion: v1
kind: PersistentVolume
metadata:
name: ollama-local-pv
spec:
capacity:
storage: 100Gi
volumeMode: Filesystem
accessModes:
- ReadWriteOnce
persistentVolumeReclaimPolicy: Retain
storageClassName: local-path
local:
path: /var/lib/ollama
nodeAffinity:
required:
nodeSelectorTerms:
- matchExpressions:
- key: kubernetes.io/hostname
operator: In
values:
- uk-desktop.tail2fe2d.ts.net

View File

@@ -3,6 +3,20 @@ image:
pullPolicy: Always
tag: "latest"
nodeSelector:
kubernetes.io/hostname: master.tail2fe2d.ts.net
kubernetes.io/hostname: uk-desktop.tail2fe2d.ts.net
tolerations:
- key: workload
operator: Equal
value: desktop
effect: NoSchedule
ingress:
enabled: false
ollama:
gpu:
enabled: true
type: 'nvidia'
number: 1
persistentVolume:
enabled: true
size: 100Gi
storageClass: "local-path"

View File

@@ -1,4 +1,4 @@
clusterDomain: ai.hexor.cy
clusterDomain: cluster.local
extraEnvVars:
GLOBAL_LOG_LEVEL: debug
@@ -32,12 +32,22 @@ ollama:
pipelines:
enabled: true
nodeSelector:
kubernetes.io/hostname: master.tail2fe2d.ts.net
tika:
enabled: true
nodeSelector:
kubernetes.io/hostname: master.tail2fe2d.ts.net
websocket:
enabled: true
nodeSelector:
kubernetes.io/hostname: master.tail2fe2d.ts.net
redis:
master:
nodeSelector:
kubernetes.io/hostname: master.tail2fe2d.ts.net
ingress:
enabled: true
@@ -46,7 +56,5 @@ ingress:
cert-manager.io/cluster-issuer: letsencrypt
traefik.ingress.kubernetes.io/router.middlewares: kube-system-https-redirect@kubernetescrd
host: "ai.hexor.cy"
tls:
- hosts:
- '*.hexor.cy'
secretName: ollama-tls
tls: true
existingSecret: ollama-tls

View File

@@ -0,0 +1,9 @@
apiVersion: apps/v1
kind: Deployment
metadata:
name: ollama
namespace: ollama
spec:
template:
spec:
runtimeClassName: nvidia

20
k8s/core/gpu/app.yaml Normal file
View File

@@ -0,0 +1,20 @@
apiVersion: argoproj.io/v1alpha1
kind: Application
metadata:
name: gpu-system
namespace: argocd
spec:
project: core
destination:
namespace: gpu-system
server: https://kubernetes.default.svc
source:
repoURL: ssh://git@gt.hexor.cy:30022/ab/homelab.git
targetRevision: HEAD
path: k8s/core/gpu
syncPolicy:
automated:
selfHeal: true
prune: true
syncOptions:
- CreateNamespace=true

View File

@@ -0,0 +1,15 @@
apiVersion: kustomize.config.k8s.io/v1beta1
kind: Kustomization
resources:
- app.yaml
- runtime-class.yaml
helmCharts:
- name: nvidia-device-plugin
repo: https://nvidia.github.io/k8s-device-plugin
version: 0.17.0
releaseName: nvidia-device-plugin
namespace: gpu-system
valuesFile: values.yaml
includeCRDs: true

View File

@@ -0,0 +1,5 @@
apiVersion: node.k8s.io/v1
kind: RuntimeClass
metadata:
name: nvidia
handler: nvidia

24
k8s/core/gpu/values.yaml Normal file
View File

@@ -0,0 +1,24 @@
nodeSelector:
kubernetes.io/hostname: uk-desktop.tail2fe2d.ts.net
tolerations:
- key: workload
operator: Equal
value: desktop
effect: NoSchedule
runtimeClassName: nvidia
setAsDefault: false
config:
name: time-slicing-config
default: any
map:
any: |-
version: v1
sharing:
timeSlicing:
resources:
- name: nvidia.com/gpu
replicas: 4