Creating a Job YAML File

Procedure

1. Download the YAML file from cluster scheduling componentGitee code repository.

   Table 1 YAML files of different job types and hardware models

   Job Type	Hardware Model	YAML File Path
   Jobs in Kubernetes or other scheduler scenarios	Atlas 200I Soc A1 core board	samples/inference/infer-310p-1usoc.yaml
   	Inference nodes of other types	samples/inference/infer.yaml
   Deployment jobs scheduled by Volcano	Atlas 200I Soc A1 core board	samples/inference/infer-deploy-310p-1usoc.yaml
   	Inference nodes of other types	samples/inference/infer-deploy.yaml

2. Upload the YAML file to any directory on the master node and modify the file content as required.

   Table 2 Parameters in the YAML file

   Parameter	Value	Description
   image	-	Inference image name. Set this parameter as required.
   replicas	Integer	replicas: number of job replicas. Generally, the value is 1.
   requests	Ascend 310 environment: huawei.com/Ascend310: number_of_processors Ascend 310P environment: huawei.com/Ascend310P: number_of_processors Example: huawei.com/Ascend310: 1	Type and number of requested NPUs. Change them as required. For these two parameters, the processor names and quantity must be the same.
   limits
   host-arch	ARM environment: huawei-arm x86 environment: huawei-x86	Architecture of the node where an inference job is executed. Set this parameter as required. The Atlas 200I Soc A1 core board node supports only huawei-arm.
   servertype	soc	Server type To schedule jobs to the Atlas 200I Soc A1 core board node, add this parameter and mount the directory by referring to the infer-310p-1usoc.yaml file. This parameter is not required for other types of nodes.

   The following uses infer-310p-1usoc.yaml as an example to describe how to set parameters.

   apiVersion: batch/v1
   kind: Job
   metadata:
     name: resnetinfer1-1-1usoc
   spec:
   ...
         nodeSelector:
           host-arch: huawei-arm
           servertype: soc
         containers:
         - image: ubuntu-infer:v1
   ...
           resources:
             requests:
               huawei.com/Ascend310P: 1
             limits:
               huawei.com/Ascend310P: 1
   ...

3. The directories and files to be mounted to the Atlas 200I Soc A1 core board node are different from those to other types of nodes. To avoid inference failure, if the Ascend 310P processor is required and the Atlas 200I Soc A1 core board node exists in the cluster but you do not want to schedule the job to this type of node, add the affinity field to the example YAML file, indicating not to schedule jobs to the nodes with the servertype=soc label.
   • The following gives an example of infer.yaml, which is the YAML file for a job.

     apiVersion: batch/v1
     kind: Job
     metadata:
       name: resnetinfer1-1
     spec:
       template:
         spec:
           affinity:
             nodeAffinity:
               requiredDuringSchedulingIgnoredDuringExecution:
                 nodeSelectorTerms:
                   - matchExpressions:
                       - key: servertype
                         operator: NotIn
                         values:
                           - soc
           nodeSelector:
             host-arch: huawei-arm 
     ...

   • The following gives an example of infer-deployment.yaml, which is the YAML file for a Deployment job.

     apiVersion: apps/v1
     kind: Deployment
     ...
     spec:
       template:
         metadata: 
           labels:
              app: infers
     ...
         spec:
           affinity:
             nodeAffinity:
               requiredDuringSchedulingIgnoredDuringExecution:
                 nodeSelectorTerms:
                   - matchExpressions:
                       - key: servertype
                         operator: NotIn
                         values:
                           - soc
           schedulerName: volcano 
           nodeSelector:
             host-arch: huawei-arm 
     ...