By using AWS re:Post, you agree to the AWS re:Post Terms of Use
AWS re:Postre:Post

EMR clusters not running

0

Hi! I had several workflows that are running on AWS EMR with spark. Since a yesterday the cluster creation step is failing.

On the events of the cluster I see a VALIDATION_ERROR with no details. On the termination message I see EMR_DefaultRole has insufficient EC2 permissions

I've recheck my roles, recreate them, and also give them the admin policy and even with all of this the cluster does not start. Also I thought it could be some aws bug deployed so I've tried different combinations of instance type, emr versions, creating new default roles, etc. Nothing worked. On cloud trail I always see that the role is not authorized on ec2, even with the full admin policy.

Has someone experience something similar? how did you fix it? Thanks in advance!

Topics
Security, Identity, & ComplianceAnalytics
Tags
AWS Identity and Access ManagementAmazon EMR
Language
English
Hector
asked a year ago293 views
1 Answer
0
Accepted Answer

Check to see if you have a policy or SCP in place that requires the Instance Metadata Service V2 that is being applied to the account you are running in. This can prevent starting any instances that are not at IMDS v2. This shows as a permissions error and is really hard to link back to a policy issue/change. https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/configuring-instance-metadata-service.html

profile pictureAWS
Keith_AWS
answered a year ago
AWS
SUPPORT ENGINEER
Yokesh NK
reviewed a month ago
  • Hector
    10 months ago

    Thanks! I found it was a restriction with the SCP that did not allow starting instances on subnets with public IPs. Changing it to private ones solved the issue

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.

Guidelines for Answering Questions

Relevant content