mirror/cube-studio
2
0
Fork 0
mirror of https://github.com/tencentmusic/cube-studio.git synced 2024-11-21 01:16:33 +08:00
Code Issues Packages Projects Releases Wiki Activity

添加spark serverless支持

Browse Source
This commit is contained in:
pengluan 2022-07-18 16:09:38 +08:00
parent 92909f194e
commit b37be5a4da
2 changed files with 7 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
install/kubernetes/start.sh
View File

@ -91,7 +91,7 @@ kubectl apply -f gpu/nvidia-device-plugin.yml
kubectl apply -f gpu/dcgm-exporter.yaml
kubectl apply -f gpu/dcgm-exporter-sm.yaml
# 部署frameworkcontroller
# 部署frameworkcontroller nni超参搜索使用
kubectl create serviceaccount frameworkcontroller --namespace kubeflow
kubectl create clusterrolebinding frameworkcontroller-kubeflow --clusterrole=cluster-admin --user=system:serviceaccount:kubeflow:frameworkcontroller
kubectl create -f frameworkcontroller/frameworkcontroller-with-default-config.yaml
@ -129,6 +129,11 @@ kubectl wait crd/applications.app.k8s.io --for condition=established --timeout=6
kubectl apply -k env/platform-agnostic
cd ../../../../
# 部署trainjob:tfjob/pytorchjob/mpijob/mxnetjob/xgboostjobs
kubectl apply -k kubeflow/train-operator/manifests/overlays/standalone
# 部署sparkjob
kubectl apply -f spark/install.yaml
# 部署管理平台
kubectl delete configmap kubernetes-config -n infra

2
job-template/job/pytorch_distributed_train_k8s/Dockerfile
View File

@ -33,6 +33,6 @@ COPY job/pkgs /app/job/pkgs
WORKDIR /app
ENV PYTHONPATH=/app:$PYTHONPATH
ENTRYPOINT ["python3", "pytorchjob_launcher.py"]
ENTRYPOINT ["python3", "launcher.py"]