我在 kubernetes 集群中的一个 GPU 工作节点上运行了一个 Nvidia deepstream 容器,它被部署为一个作业,下面是我的 YAML 文件,
apiVersion: batch/v1
kind: Job
metadata:
name: deepstream-test
spec:
backoffLimit: 1
template:
spec:
containers:
- name: nvidia-deepstream
image: lkkrishna945/deepstream-5.0:test
command: ["deepstream-app","-c","/opt/nvidia/deepstream/deepstream-5.0/samples/configs/deepstream-app/source30_1080p_dec_infer-resnet_tiled_display_int8_edited.txt"]
ports:
- containerPort: 8554
resources:
limits:
nvidia.com/gpu: 1
restartPolicy: Never
在我部署此作业后,它运行良好,但其输出运行在 on-perm 开源 kubernetes 集群的一个工作节点上,但我想流式传输作为视频的运行输出。
这是我使用基础 Nvidia deepstream 容器构建的 Dockerfile,
FROM nvcr.io/nvidia/deepstream:5.0-dp-20.04-triton
ADD source30_1080p_dec_infer-resnet_tiled_display_int8_edited.txt /opt/nvidia/deepstream/deepstream-5.0/samples/configs/deepstream-app/
CMD ["deepstream-app -c /samples/configs/deepstream-app/source30_1080p_dec_infer-resnet_tiled_display_int8_edited.txt"]
任何人都可以对此提供任何建议/解决方案吗?