run_kubernetes

You have to make sure that there is mounted Azure File Storage on the machine you are runnning this code.

1. df -h to check mounted AFS

2. AFS has to be mounted to the /mnt/shipfs folder and have enough space

3. Make sure yout config.py is correct:

   K8S_PROXY - check if you do not have a connection to kubernetes (kubectl get jobs)

   HOST_OUTPUT_DIRECTORY - set the output folder for your experiments

   DOCKER_OUTPUT_DIRECTORY, DOCKER_SAMPLE_DIRECTORY - set these parameters according to your docker image

   TIMEOUT - set your timeout for every Job

   JOB_SPEC - check this

4. Launch - python run_kubernetes.py

5. Every Kubernetes Job create 1 Kubernetes Pod. Monitor your Jobs with kubectl utility:
   <pod_name> = <job_name>-[:5]

kubectl --server=127.0.0.1:8002 get jobs      
kubectl --server=127.0.0.1:8002 get pods    
kubectl --server=127.0.0.1:8002 describe pod <pod_name>    
kubectl --server=127.0.0.1:8002 logs <pod_name>

run_ship_at

scripts to run fs