# CANN应用类日志存储至OBS配置方法
#### 创建OBS的PVC
1. 创建OBS桶。
2. 在CCE控制台创建关联OBS桶的PV及PVC。
   1. 进入要挂载OBS的CCE集群控制台。 ![](https://support.huaweicloud.com/bestpractice-modelarts/figure/zh-cn_image_0000002503500623.png "点击放大")
      
   
   2. 创建存储卷PV。 ![](https://support.huaweicloud.com/bestpractice-modelarts/figure/zh-cn_image_0000002503500619.png "点击放大")
      ![](https://support.huaweicloud.com/bestpractice-modelarts/figure/zh-cn_image_0000002470740662.png "点击放大")
      
   
   3. 创建存储卷声明pvc ![](https://support.huaweicloud.com/bestpractice-modelarts/figure/zh-cn_image_0000002503660541.png "点击放大")
      
    
3. 创建完成后，在CCE中可以看到对应的PVC，在容器中直接引用即可。
   ```
   kubectl get pvc pvc-obs-glj -o yaml
   ```
   ![](https://support.huaweicloud.com/bestpractice-modelarts/figure/zh-cn_image_0000002503660535.png "点击放大")
   
 
#### 训练作业关联PVC
若需要将日志写入对应的桶中，通常需要对训练的[config.yaml](https://support.huaweicloud.com/bestpractice-modelarts/modelarts_llm_train_c_5910008.html)做如下改造以引用上一步创建的PVC。
- 获取jobid和nodename以用于将每个任务产生的日志写入特定的桶中 ![](https://support.huaweicloud.com/bestpractice-modelarts/figure/zh-cn_image_0000002470580672.png)
  
- 定义一个重定向的路径，引用上一步取到的jobid 和nodename ![](https://support.huaweicloud.com/bestpractice-modelarts/figure/zh-cn_image_0000002503500625.png "点击放大")
  

- 定义转储规则 ![](https://support.huaweicloud.com/bestpractice-modelarts/figure/zh-cn_image_0000002470740658.png)
  

- 一个完整的yaml示例（供参考）
  ```
  apiVersion: batch.volcano.sh/v1alpha1
  kind: HyperJob 
  metadata: 
    name: hyperjob-test
  spec:
    replicatedJobs:
     - replicas: 2 # vcjob的数量，和后面启动命令的WORD_SIZE及RANK有联动
       name: vcjob-test # vcjob的名称
       template:WORD_SIZE
         minAvailable: 2 # vcjob内最小可用数量，AI任务场景等于每个vcjob内副本数
         tasks:
          - replicas: 2 # Pod的副本数，和后面启动命令的WORD_SIZE及RANK有联动
            name: worker  # Pod的名称
            policies:
            - event: PodEvicted
              action: RestartJob
            template:
              spec:
                volumes:
                  - name: ascend-driver
                    hostPath: {path: /usr/local/Ascend/driver}
                  - name: ascend-add-ons
                    hostPath: {path: /usr/local/Ascend/add-ons}
                  - name: npu-smi
                    hostPath: {path: /usr/local/sbin/npu-smi}
                  - name: localtime
                    hostPath: {path: /etc/localtime}
                  - name: sfs
                    hostPath: {path: /mnt/sfs_turbo/}
                  - name: ascend-install
                    hostPath: {path: /etc/ascend_install.inf}
                  - name: dcmi
                    hostPath: {path: /usr/local/dcmi}
                  - name: pvc-obs-glj  # pvc挂载信息
                    persistentVolumeClaim:
                      claimName: pvc-obs-glj
                containers:
                - name: ${container_name} # 容器名称
                  image: ${image_name} # 镜像地址
                  command: ["/bin/bash", "-c"]
                  args:
                    - ${command}
                  resources:
                    requests: 
                      huawei.com/ascend-1980: "16"  # 申请的NPU数量，需与limits值保持一致，系统将根据申请NPU的数量决定Pod是否独占节点
                    limits: 
                      huawei.com/ascend-1980: "16" # 修改项 每个节点的需求卡数，key保持不变。
                  ports:
                  - containerPort: 29500
                    name: trainport
                    protocol: TCP
                  volumeMounts:
                  - name: ascend-driver # 驱动挂载，保持不动
                    mountPath: /usr/local/Ascend/driver
                  - name: ascend-add-ons # 驱动挂载，保持不动
                    mountPath: /usr/local/Ascend/add-ons
                  - name: npu-smi
                    mountPath: /usr/local/sbin/npu-smi
                  - name: localtime
                    mountPath: /etc/localtime
                  - name: sfs
                    mountPath: /mnt/sfs_turbo/
                  - name: ascend-install
                    mountPath: /etc/ascend_install.inf
                  - name: dcmi
                    mountPath: /usr/local/dcmi
  - name: pvc-obs-glj
                    mountPath: /data/logs
                  env:
  - name: NODE_NAME
                    valueFrom:
                      fieldRef:
                        apiVersion: v1
                        fieldPath: spec.nodeName
                  - name: JOB_NAME
                    valueFrom:
                      fieldRef:
                        apiVersion: v1
                        fieldPath: metadata.labels['app']
                  - name: ASCEND_PROCESS_LOG_PATH
                    value: /data/logs/$(JOB_NAME)/$(NODE_NAME)/ascend_plog/
                  - name: ASCEND_WORK_PATH
                    value: /data/logs/$(JOB_NAME)/$(NODE_NAME)/ascend_work_path/log/
                  - name: HYPERJOB_NAME
                    valueFrom:
                      fieldRef:
                        fieldPath: metadata.annotations['volcano.sh/hyperjob-name']
                  - name: HYPERJOB_REPLICATEDJOB_NAME
                    valueFrom:
                      fieldRef:
                        fieldPath: metadata.annotations['volcano.sh/hyperjob-replicatedjob-name']
                  - name: TASK_SPEC
                    valueFrom:
                      fieldRef:
                        fieldPath: metadata.annotations['volcano.sh/task-spec']
                  - name: HYPERJOB_REPLICATEDJOB_INDEX
                    valueFrom:
                      fieldRef:
                        fieldPath: metadata.annotations['volcano.sh/hyperjob-replicatedjob-index']
                  - name: TASK_INDEX
                    valueFrom:
                      fieldRef:
                        fieldPath: metadata.annotations['volcano.sh/task-index']
                restartPolicy: OnFailure
  ```
  
如何启动一个带独立uuid的作业：
```
kubectl apply -f <(sed "s/pytorch-dist-git2/pytorch-dist-git2-$(uuidgen | head -c 6)/g" pytorchjob-qwen2.5VL-3B-lora.yaml）
```