Chapter 42: Terraform Drift and Troubleshooting
Chala mitrano, aaj drift aani troubleshooting. Koni console madhe badal kela tar code aani AWS vegle hotat — tyala drift mhantat. Roj ratri plan chalavun te pakdaycha. Ani error aali tar ghabraycha nahi: pahili line vacha, TF_LOG lava, aani table madhe shodha.
What you will learn in this chapter
- What drift is and how a nightly plan detects it
- Codify or revert, never click again
- Debug logs with TF_LOG and TF_LOG_PATH
- A runbook table of common errors
Concepts in this chapter
- 42.1Drift — when AWS no longer matches the code
- 42.2Troubleshooting with TF_LOG and the common errors
- 42.3Troubleshooting table — error, cause, fix
The chapter recap is at the end of the last concept page.
Practice Tasks
Suppose we are on the DevOps team at BBC, the news media company, working from London. Oliver Hughes leads the team, and we look after the live news ticker. Costs are counted in GBP. In the tasks below we do real DevOps jobs on this project using what this lesson teaches. BBC is a real company, but this project, the people, servers, IP addresses and numbers are made up for practice.
0 / 12
Easy Name the practice
Why: IaC is the base of repeatable cloud setups.
What to do: BBC describes its servers, networks and databases in text files kept in Git and creates them with a tool. What is this called (three-letter short form)?
- Write down the facts or numbers the task gives you.
- Work it out step by step.
- Type only the final answer (one word or number) in the box.
- Click Check answer.
Easy S3 bucket resource
Why: Buckets in code are reviewed and never forgotten.
What to do: Write a resource aws_s3_bucket named logs with bucket = "bbc-logs" and tags { Team = "devops" }.
- Write the block type (resource, variable, output, module...) and its labels in quotes.
- Open { and write each argument as name = value.
- Put text values in double quotes; numbers and true/false without quotes.
- Close every block with } and click Check answer.
Easy Output a value
Why: Outputs hand values to scripts and other modules.
What to do: Write output "web_ip" with value = aws_instance.web.public_ip.
- Write the block type (resource, variable, output, module...) and its labels in quotes.
- Open { and write each argument as name = value.
- Put text values in double quotes; numbers and true/false without quotes.
- Close every block with } and click Check answer.
Medium List resources in state
Why: state list shows what Terraform thinks it manages.
What to do: List every resource address in the state.
- Read the task and pick the tool it needs (git, docker, kubectl, terraform, ansible...).
- Write the tool name, then the sub-command (the action word).
- Add the flags the task needs, each followed by its value.
- Add names, files or paths at the end, then click Check answer.
Medium Save a plan file
Why: Applying a saved plan applies exactly what was reviewed.
What to do: Create a plan and save it to the file bbc.tfplan.
- Read the task and pick the tool it needs (git, docker, kubectl, terraform, ansible...).
- Write the tool name, then the sub-command (the action word).
- Add the flags the task needs, each followed by its value.
- Add names, files or paths at the end, then click Check answer.
Medium Plan with a variable value
Why: The same code builds dev and prod with different values.
What to do: Run a plan with the variable region set to eu-west-2 and using the file prod.tfvars.
- Read the task and pick the tool it needs (git, docker, kubectl, terraform, ansible...).
- Write the tool name, then the sub-command (the action word).
- Add the flags the task needs, each followed by its value.
- Add names, files or paths at the end, then click Check answer.
Medium EC2 instance using a variable
Why: Referencing var.* avoids hard-coding sizes.
What to do: Write resource aws_instance "web" with ami = "ami-0abcdef1234567890", instance_type = var.instance_type and tags Name = "live-news".
- Write the block type (resource, variable, output, module...) and its labels in quotes.
- Open { and write each argument as name = value.
- Put text values in double quotes; numbers and true/false without quotes.
- Close every block with } and click Check answer.
Hard Remote state in S3
Why: Shared remote state with locking stops two people breaking each other.
What to do: Write a terraform block with backend "s3": bucket = "bbc-tfstate", key = "prod/terraform.tfstate", region = "eu-west-2", dynamodb_table = "tf-lock".
- Write the block type (resource, variable, output, module...) and its labels in quotes.
- Open { and write each argument as name = value.
- Put text values in double quotes; numbers and true/false without quotes.
- Close every block with } and click Check answer.
Hard Import an existing bucket
Why: Import brings hand-made resources under code control.
What to do: Import the existing S3 bucket bbc-logs into the address aws_s3_bucket.logs.
- Read the task and pick the tool it needs (git, docker, kubectl, terraform, ansible...).
- Write the tool name, then the sub-command (the action word).
- Add the flags the task needs, each followed by its value.
- Add names, files or paths at the end, then click Check answer.
Hard Plan one resource only
Why: During an incident you may need to touch only one resource.
What to do: Plan only the resource aws_instance.web.
- Read the task and pick the tool it needs (git, docker, kubectl, terraform, ansible...).
- Write the tool name, then the sub-command (the action word).
- Add the flags the task needs, each followed by its value.
- Add names, files or paths at the end, then click Check answer.
Hard Create several copies with count
Why: count builds N identical resources from one block.
What to do: Write resource aws_instance "worker" with count = 2, ami = "ami-0abcdef1234567890", instance_type = "t3.small" and tags Name = "worker-${count.index}".
- Write the block type (resource, variable, output, module...) and its labels in quotes.
- Open { and write each argument as name = value.
- Put text values in double quotes; numbers and true/false without quotes.
- Close every block with } and click Check answer.
Hard Call a module
Why: Modules package a tested pattern, like a VPC, for reuse.
What to do: Write module "vpc" with source = "terraform-aws-modules/vpc/aws", version = "5.8.1", name = "bbc-vpc" and cidr = "10.0.0.0/16".
- Write the block type (resource, variable, output, module...) and its labels in quotes.
- Open { and write each argument as name = value.
- Put text values in double quotes; numbers and true/false without quotes.
- Close every block with } and click Check answer.
Ravi Sir's Cards: 9 revision, interview and comparison cards for this chapter · मराठी cards