Custom Resource Type creation fails with Internal Failure
Closed this issue · 2 comments
Working through the Workshop on creating Resource Types using the python and typescript examples in this repository, the Resource Types are registered successfully, but attempting to use them in a CloudFormation Stack always results in an Internal Failure.
I realize this isn't necessarily an issue with the samples, hoping to gain some insight on possible causes from the team given that these samples are assumed to function.
Additional Info:
- The
Internal Failure
appears to occur before the handler code is loaded or executed as it happens immediately and there are no CloudWatch Logs created - I am able to locate and see (in CloudTrail) the AssumeRole that CloudFormation makes to the ExecutionRole created by the
cloudformation-cli
- It was suggested on an internal thread that building/deploying the Resource Type from OSX may have caused an issue with the container image, so have shifted to building/deploying from CodeBuild
- This
Internal Failure
occurs on all Resource Types I have registered, even those created manually - Bucket level ACLs are not allowed at the Org level for my accounts, so I am using a forked version of the
cloudformation-cli
based off of version 0.2.28 where I have removed the canned ACLs for the S3 Buckets created by theCloudFormationManagedUploadInfrastructure
stack
Hi @chamcca - if you upgrade your copy of the CloudFormation CLI to the latest version at this time (0.2.29
), and if you do the same for your copy of the cloudformation-cli-python-plugin
, do you still experience issues, or some of them? For the latter, which specific issue do you see?
upgrading to the new version of the cli (0.2.29) eliminated the need to use a forked version (the canned ACL appears to be removed from the LogsBucket in that stack). subsequent deployments of the python and typescript examples appear to be working this morning. I believe I am unblocked.