Pod template File not found exception though it is exist

2

Hello, I have a driver and executor pod template files exist in s3. But when I run the job, it failed wirh FileNotFoundException: s3:/<s3bucket>/podtemplate/driver.yaml (no such file or directory) error.

Appreciate the help and suggestions here. Thank in advance

Mark
已提問 6 個月前檢視次數 233 次
1 個回答
3
已接受的答案

Hello,

There could a couple of reasons as mentioned below that might cause this issue,

  1. Try to use local file system directory to pass the pod template and see if that works.
  2. Check the execution role arn that submits the spark which has sufficient permission to read the pod file from S3. More details here -

https://docs.aws.amazon.com/emr/latest/EMR-on-EKS-DevelopmentGuide/creating-job-execution-role.html &

https://docs.aws.amazon.com/emr/latest/EMR-on-EKS-DevelopmentGuide/iam-execution-role.html

  1. Check if the mentioned S3 bucket has any bucket policy restriction that might block the read access. https://docs.aws.amazon.com/AmazonS3/latest/userguide/example-bucket-policies.html
  2. You can leverage IAM simulator to test if the given policy can be accessible to s3 pod files - https://docs.aws.amazon.com/IAM/latest/UserGuide/access_policies_testing-policies.html
  3. Test the sample pod template job given this doc which helps to see if the issue with all the pod file template or only with problematic one - https://catalog.us-east-1.prod.workshops.aws/workshops/1f91e1d4-5587-40ff-8d5d-54fc86e0ddc1/en-US/advanced/pod-template
  4. Enable and leverage the control plane, audit logs and manager logs to see if anything reported - https://docs.aws.amazon.com/eks/latest/userguide/control-plane-logs.html
  5. Run this command before starting their testing. It will capture all events triggered by k8s during their testing. This will help to find if any issue in the events that initiated when launching the spark job.

kubectl get ev -n <namespace used in EMR virtual cluster> -w

AWS
支援工程師
已回答 6 個月前
profile pictureAWS
專家
已審閱 6 個月前
  • Thank you for the detailed answer. With the help of your suggestion, I sorted the issue that occurred due to permission missing.

您尚未登入。 登入 去張貼答案。

一個好的回答可以清楚地回答問題並提供建設性的意見回饋,同時有助於提問者的專業成長。

回答問題指南