terraform.md

Terraform

• NOTES:
  • This section depends on Kubernetes and a ~/.kubeconfig from above
  • NodePort was used instead of LoadBalancer for service.type
    • MetalLB is a stretch goal for future deployments
• Install terraform via `asdf

  # terraform
  asdf plugin-add terraform
  asdf install terraform latest

• Add aliases to ~/.bashrc or ~/.zshrc

  # ~/.bashrc
  alias tf='terraform'
  alias tfi='terraform init -backend-config=./state.conf'
  alias tfa='terraform apply'
  alias tfp='terraform plan'
  alias tfpn='terraform plan -refresh=false'

• Navigate to ./terraform/ and initialize the terraform working directory

  cd terraform/
  tfi

• Create an execution plan

  tfp

• Apply/execute the actions from Terraform plan

  tfa

• Navigate to http://localhost:<port>
  • Port can be found via kubectl

    k get svc   # 80:31942/TCP

• Tear down deployment

  tf destroy

  • Real-time view of pod removal

Debugging

• Verbose logging and redirection to a file

  export TF_LOG="trace"                       # unset via "off"
  export TF_LOG_PATH="$HOME/Downloads/tf.log" # `~` doesn't expand

  • Log levels
    • TRACE
    • DEBUG
    • INFO
    • WARN
    • ERROR

• Error: cannot re-use a name that is still in use

  I think I resolved the issue. This is what I did: 1) mv the terraform.tfstate to another name, 2) mv the terraform.tfstate.backup to terraform.tfstate, and 3) run 'terraform refresh' command to confirm the state is synchronized, and 4) run 'terraform apply' to delete/create the resource. I will mark your reply as the answer, as it gives me the clue for solving the issue. Thanks! – ozmhsh Dec 9, 2021 at 4:57

  nginx - Stuck in the partial helm release on Terraform to Kubernetes - Stack Overflow