How to fix Error: Provisioner execution failed - command not found in Terraform

Replace the command name with its absolute path. For example, instead of:

provisioner "local-exec" {
  command = "gcloud auth list"
}

Use:

provisioner "local-exec" {
  command = "/usr/bin/gcloud auth list"
}

Find the full path using which <command> on Linux/macOS or where <command> on Windows.

Use the interpreter argument to ensure commands run in a known shell environment:

provisioner "local-exec" {
  interpreter = ["bash", "-c"]
  command     = "echo 'Running in bash'"
}

This is especially important when using shell-specific syntax like variable expansion or pipes.

Use the working_dir argument to ensure relative paths resolve correctly:

provisioner "local-exec" {
  command     = "./deploy.sh"
  working_dir = "${path.module}/scripts"
}

Ensure the directory exists before the provisioner runs.

If running in Terraform Cloud, ensure all required tools are available in that environment. For the gcloud example, you may need to:

provisioner "local-exec" {
  command = "curl https://sdk.cloud.google.com | bash"
}

provisioner "local-exec" {
  command = "gcloud auth list"
}

Alternatively, use a custom Docker image or build step that includes required tools.

When referencing your own scripts, use absolute paths instead of relative paths:

provisioner "local-exec" {
  command = "${path.module}/scripts/deploy.sh"
}

Never rely on the current working directory for script resolution.

Use on_failure = "continue" temporarily to see error output without stopping the entire apply:

provisioner "local-exec" {
  command     = "failing-command"
  on_failure  = "continue"
}

This helps you diagnose PATH or environment issues without rolling back resources. Once fixed, remove this argument or set it back to "fail".