To set Terraform logs to the verbose mode:
export TF_LOG=trace
To remove corrupt state:
terraform state rm module.loki.helm_release.loki-stack
To refresh tfstate:
terraform refresh -var-file example.tfvars
To recreate resources:
terraform taint module.system.null_resource.helm_init
If terraform destroy command fails, run destroy_fix.sh and try terraform destroy again. After successful destroy process go to AWS console and delete argo-artifacts S3 bucket (if needed), also delete Route53 resources remaining from your deployment.
Further reading:
Sometimes public IP address appears in blacklist of some services (e.g https://infra.apache.org/infra-ban.html). To avoid that - is good to prevent abuse such services using artifact caching servers such as Nexus, Artifactory, etc (if they have their own public IP) If there is still need to have direct access - it's time to just recreate IP.
In swiss-army-kube/example (or swiss-army-kube/<environment name>):
terraform destroy -target 'module.network.module.vpc.aws_eip.nat[0]'
terraform apply -target module.network.module.vpcThis method is suitable when short absence of access from VPC to internet is affordable (e.g during maintenance)