fatal error: concurrent map writes

[davidorlea]

Describe the bug

Since upgrading from Terragrunt 0.59.6 to 0.67.13 we are repeatedly getting fatal error: concurrent map writes when calling terragrunt init, terragrunt validate, terragrunt validate-inputs or terragrunt plan. This is not happening in every call to terragrunt, but pretty regularly, effectively rendering our CI pipelines unusable.

The stack trace is always from

github.com/gruntwork-io/terragrunt/internal/cache.(*ExpiringCache[...]).Get(0x2884e40, {0x2884628, 0xc000d41030}, {0xc000062249, 0x4c})
    /home/circleci/project/internal/cache/cache.go:99 +0x205

I attached a few examples with the complete stack traces.

We experience this locally on Apple ARM-based (M1, M2) MacBooks, as well as AWS EC2 instances running on amd64-based architecture.

With Terragrunt 0.59.6 this did not happen once to us.

Steps To Reproduce

Can't be reproduced explicitly, but occurs regularly every few terragrunt-calls.

Expected behavior

No exception and clean exit from Terragrunt.

Nice to haves

• [x ] Terminal output
• [ ] Screenshots

Versions

• Terragrunt version: 0.67.13
• Terraform version: Terraform 0.12.31, 0.13.7, 0.14.11, 0.15.5, 1.5.7
• Environment details: AWS EC2 amd64, Apple ARM M1, M2

Additional context

fatal1.txt fatal2.txt fatal3.txt

[yhakbar]

Thanks for reporting this, @davidorlea . We'll look into it soon.

[yhakbar]

Fix pre-released in v0.68.0-beta2024100702. Will be included in the next minor release.

[davidorlea]

Very much appreciate your quick response ❤️

I can confirm that it's running smoothly now!