Merge branch 'add_example' into 'master'

[Issue volcano-sh#49] 添加 openmpi 及 tensorflow 的 job样例 ![image](/uploads/132c7cb1612d13b6f6ab40855eee77f1/image.png) ![image](/uploads/3b09ea556e1acabfe79e59ec15af08a2/image.png) Issues info: Issue ID: 49 Title: 添加 openmpi 及 tensorflow 的 job样例 Issue url: CBU-PaaS/Community/volcano/volcano#49 See merge request CBU-PaaS/Community/volcano/volcano!81
wangyuqing4 · Apr 1, 2019 · 30dcab8 · 30dcab8
2 parents 549d7e7 + 8a413fa
commit 30dcab8
Show file tree

Hide file tree

Showing 2 changed files with 117 additions and 0 deletions.
diff --git a/example/openmpi-hello.yaml b/example/openmpi-hello.yaml
@@ -0,0 +1,55 @@
+apiVersion: batch.volcano.sh/v1alpha1
+kind: Job
+metadata:
+  name: openmpi-hello
+spec:
+  minAvailable: 3
+  schedulerName: kube-batch
+  plugins:
+    ssh: []
+    env: []
+  tasks:
+  - replicas: 1
+    name: mpimaster
+    policies:
+      - event: TaskCompleted
+        action: CompleteJob
+    template:
+      spec:
+        imagePullSecrets:
+          - name: default-secret
+        containers:
+        - command:
+            - /bin/sh
+            - -c
+            - |
+              MPI_HOST=`cat /etc/volcano/mpiworker.host | tr "\n" ","`;
+              mkdir -p /var/run/sshd; /usr/sbin/sshd;
+              mpiexec --allow-run-as-root --host ${MPI_HOST} -np 2 mpi_hello_world > /home/re
+          image: 100.125.5.235:20202/l00427178/openmpi-hello:3.28
+          name: mpimaster
+          ports:
+          - containerPort: 22
+            name: mpijob-port
+          workingDir: /home
+        restartPolicy: OnFailure
+  - replicas: 2
+    name: mpiworker
+    template:
+      spec:
+        imagePullSecrets:
+          - name: default-secret
+        containers:
+        - command:
+            - /bin/sh
+            - -c
+            - |
+              mkdir -p /var/run/sshd; /usr/sbin/sshd -D;
+          image: 100.125.5.235:20202/l00427178/openmpi-hello:3.28
+          name: mpiworker
+          ports:
+          - containerPort: 22
+            name: mpijob-port
+          workingDir: /home
+        restartPolicy: OnFailure
+
diff --git a/example/tensorflow-benchmark.yaml b/example/tensorflow-benchmark.yaml
@@ -0,0 +1,62 @@
+apiVersion: batch.volcano.sh/v1alpha1
+kind: Job
+metadata:
+  name: tensorflow-benchmark
+spec:
+  minAvailable: 5
+  schedulerName: kube-batch
+  plugins:
+    env: []
+  policies:
+  - event: PodEvicted
+    action: RestartJob
+  - event: PodFailed
+    action: RestartTask
+  tasks:
+  - replicas: 2
+    name: ps
+    template:
+      spec:
+        imagePullSecrets:
+          - name: default-secret
+        containers:
+        - command:
+          - sh
+          - -c
+          - |
+            PS_HOST=`cat /etc/volcano/ps.host | sed 's/$/&:2222/g' | tr "\n" ","`;
+            WORKER_HOST=`cat /etc/volcano/worker.host | sed 's/$/&:2222/g' | tr "\n" ","`;
+            python tf_cnn_benchmarks.py --batch_size=32 --model=resnet50 --variable_update=parameter_server --flush_stdout=true --num_gpus=1 --local_parameter_device=cpu --device=cpu --data_format=NHWC --job_name=ps --task_index=${VK_TASK_INDEX} --ps_hosts=${PS_HOST} --worker_hosts=${WORKER_HOST}
+          image: 100.125.5.235:20202/l00427178/tf-benchmarks-cpu:v20171202-bdab599-dirty-284af3
+          name: tensorflow
+          ports:
+          - containerPort: 2222
+            name: tfjob-port
+          resources: {}
+          workingDir: /opt/tf-benchmarks/scripts/tf_cnn_benchmarks
+        restartPolicy: OnFailure
+  - replicas: 3
+    name: worker
+    policies:
+      - event: TaskCompleted
+        action: CompleteJob
+    template:
+      spec:
+        imagePullSecrets:
+          - name: default-secret
+        containers:
+        - command:
+          - sh
+          - -c
+          - |
+            PS_HOST=`cat /etc/volcano/ps.host | sed 's/$/&:2222/g' | tr "\n" ","`;
+            WORKER_HOST=`cat /etc/volcano/worker.host | sed 's/$/&:2222/g' | tr "\n" ","`;
+            python tf_cnn_benchmarks.py --batch_size=32 --model=resnet50 --variable_update=parameter_server --flush_stdout=true --num_gpus=1 --local_parameter_device=cpu --device=cpu --data_format=NHWC --job_name=worker --task_index=${VK_TASK_INDEX} --ps_hosts=${PS_HOST} --worker_hosts=${WORKER_HOST}
+          image: 100.125.5.235:20202/l00427178/tf-benchmarks-cpu:v20171202-bdab599-dirty-284af3
+          name: tensorflow
+          ports:
+          - containerPort: 2222
+            name: tfjob-port
+          resources: {}
+          workingDir: /opt/tf-benchmarks/scripts/tf_cnn_benchmarks
+        restartPolicy: OnFailure