kind: Deployment
apiVersion: apps/v1
metadata:
name: inference-abcccae1-94fd-4d7c-8e8d-376e6948e601
namespace: topsec-poc
spec:
replicas: 1
selector:
matchLabels:
inference: inference-abcccae1-94fd-4d7c-8e8d-376e6948e601
template:
metadata:
labels:
inference: inference-abcccae1-94fd-4d7c-8e8d-376e6948e601
spec:
volumes:
- name: shm-vol
emptyDir:
medium: Memory
sizeLimit: 20Gi
- name: container-default-volume
persistentVolumeClaim:
claimName: nfs-pvc
containers:
- name: container1
image: '10.30.15.56:5000/intelligent/comfyui:2.9.0-cuda12.8-cudnn9-runtime'
command:
- /bin/sh
args:
- '-c'
- |
python3 /workspace/ComfyUI/main.py --listen 0.0.0.0 --port 8188 &
until curl -sf curl -sf http://127.0.0.1:8188/; do
sleep 1
done
exec /gateway \
--listen :27600 \
--comfy http://127.0.0.1:8188 \
--assets /workspace/ComfyUI/assets/Z-Image-Turbo
env:
- name: MODELSCOPE_CACHE
value: /home/data/modelscope
resources:
limits:
nvidia.com/gpu: '1'
requests:
nvidia.com/gpu: '1'
volumeMounts:
- name: shm-vol
mountPath: /dev/shm
- name: container-default-volume
mountPath: /workspace/ComfyUI/models
subPath: ComfyUI/models
- name: container-default-volume
mountPath: /workspace/ComfyUI/assets
subPath: ComfyUI/assets
terminationMessagePath: /dev/termination-log
terminationMessagePolicy: File
imagePullPolicy: IfNotPresent
restartPolicy: Always
terminationGracePeriodSeconds: 30
dnsPolicy: ClusterFirst
nodeSelector:
gpu_core_mode: whole
securityContext: {}
schedulerName: default-scheduler
runtimeClassName: nvidia
strategy:
type: Recreate
revisionHistoryLimit: 10
progressDeadlineSeconds: 600
