Question

I am trying to deploy a Terraform stack from a Linux EC2 and getting the following error:

 The plugin.(*GRPCProvider).ApplyResourceChange request was cancelled.

and

│ The plugin.(*GRPCProvider).PlanResourceChange request was cancelled.

and

│ The plugin.(*GRPCProvider).ValidateResourceConfig request was cancelled.

for random resources.

Terraform versions: 1.3.1_linux_amd64, 1.3.5_linux_amd64. The reason am noting the Linux ec2 its because when depolying from my local computer I am not getting any errors, with the same Terraform versions but for mac.

provider.tf:

terraform {
  required_providers {
    aws = {
      source  = "hashicorp/aws"
      version = "~> 4.0"
    }
  }
}

provider "aws" {
  region                   = "us-east-1"
}

Answer 1

I fixed the version to 4.33, the last one that worked for me, and got rid of this error. provider.tf:

terraform {
  required_providers {
    aws = {
      source  = "hashicorp/aws"
      version = "4.33.0"
    }
  }
}

Answer 2

The AWS provider is crashing for some reason and in my case it was the atlantis pod reaching the k8s memory limit. Version 4.33 worked because it consumes less memory for the same plan (350MB vs 430MB in newer versions)

So i recommend checking your memory limits, even if the pod is not being killed, the aws provider itself is crashing/detecting the memory limit and exit. If not inside a docker, check the host memory usage

友情链接