-
I try to run privategpt in a kubernetes cluster as deployment, alongside ollama. Ollama and OpenWebUI are up and running. But I've struggles with privateGPT. Problem:: I've found this solution/workaround: #1876, which in my understanding change the uid to the uid of the host system to create the cache dir, which is not possible since kubernetes assigns a random uid to the pod. I've changed the ownership in Dockerfile with a @minixxie: I've found your question regarding running multiple replicas, how do you manage to run privateGPT in kubernetes? below my current deployment: kind: Deployment
apiVersion: apps/v1
metadata:
name: privategpt
labels:
app: privategpt
spec:
replicas: 1
selector:
matchLabels:
app: privategpt
template:
metadata:
labels:
app: privategpt
networkrole: ingress
spec:
imagePullSecrets:
- name: pull-secret
containers:
- name: privategpt
image: privategpt-image
resources:
requests:
cpu: "500m"
memory: "1Gi"
limits:
cpu: "1000m"
memory: "2Gi"
nvidia.com/gpu: "0"
envFrom:
- configMapRef:
name: privategpt-config
ports:
- containerPort: 8001
readinessProbe:
tcpSocket:
port: 8001
timeoutSeconds: 10
periodSeconds: 20
successThreshold: 1
failureThreshold: 3
livenessProbe:
tcpSocket:
port: 8001
timeoutSeconds: 5
periodSeconds: 10
successThreshold: 1
failureThreshold: 3
volumeMounts:
- name: privategpt-volume
mountPath: /home/worker/app/local_data
tty: true
volumes:
- name: privategpt-volume
persistentVolumeClaim:
claimName: privategpt-volume
``` |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment
-
Created a own docker image which uses group permission since the uid in kubernetes is random |
Beta Was this translation helpful? Give feedback.
Created a own docker image which uses group permission since the uid in kubernetes is random