INFRASTRUCTURE LEARNINGINFRASTRUCTURE LEARNING
インフラ大全インフラ大全IT INFRASTRUCTURE / LEARNING
Infrastructure learning guide
Find a lesson or lab/

Esc to close · Search terms stay on your device

LEARNING日本語

Topic lessonTopic: Account protection and resource cleanup

Stop vs terminate an EC2 instance: restartability, data retention and costs

Stop changes an instance’s state so it can later be started; terminate permanently removes the instance. Neither operation means that all data and associated charges behave identically. Use a typical EBS-backed instance to compare these actions with reboot and hibernation.

What you will learn

Choose between stopping and terminating for a stated purpose, and identify the data, EBS settings and retained-resource checks needed beforehand.

Prerequisite

EBS is a storage resource separate from the instance. This lesson performs no real stop/deletion actions or commands.

Decide whether to resume it or remove the instance

Compare normal stop and termination from running state
Compare normal stop and termination from running state1: Running — Normal stop: an instance intended for later use can start again through pending/running. — Stopped. 2: Running — Terminate: permanently remove the instance; do not treat it as a routine pause. — Terminated12RunningStoppedTerminatedCompare normal stop and termination from running state1: Running — Normal stop: an instance intended for later use can start again through pending/running. — Stopped. 2: Running — Terminate: permanently remove the instance; do not treat it as a routine pause. — Terminated12RunningStoppedTerminated
  1. Running → StoppedNormal stop: an instance intended for later use can start again through pending/running.
  2. Running → TerminatedTerminate: permanently remove the instance; do not treat it as a routine pause.
Running
Active compute; idle time can still incur compute charges.
Stopped
An EBS-backed instance can be started later; retained storage costs are separate.
Terminated
That instance cannot be started or recovered; review data settings before the action.
Main differences between a normal stop and termination
AreaNormal stopTerminate
Start the same instance againPossibleCannot start or recover that instance
EBSRetained; storage costs are separateEach volume’s DeleteOnTermination governs its retention/deletion
Instance-store dataLostLost
RAM contentsNot retainedNot retained
Associated resources/commitmentsInspect EBS, addresses, other components and commitments separatelyInspect retained resources and commitments separately

Data survival depends on where the data is stored

An EBS volume differs from temporary instance store storage on the host. A normal stop retains EBS, but loses instance-store data and RAM contents. Do not assume that stopping preserves every part of an application’s working state.

For termination, inspect DeleteOnTermination per EBS volume: true selects deletion with termination; false retains the volume. Defaults depend on creation and attachment details, so “root” or “data” labels alone are not sufficient. Review charges for retained volumes separately.

Reboot and hibernation are not normal stop

Distinguish lifecycle operations by their purpose
OperationDistinction
RebootRestart the OS on the same host; instance-store data remains and compute charges continue
Normal stopStop EBS-backed compute without retaining RAM or instance-store data
HibernateOn supported configurations, save RAM to EBS; stopping-state compute charges apply
TerminatePermanently remove the instance; EBS retention/deletion settings matter

A shutdown command inside the OS does not universally mean normal stop. For EBS-backed instances, check settings such as InstanceInitiatedShutdownBehavior. Do not infer the data or cost outcome from the command’s name alone.

Review reconnection addresses and retained charges

A stop/start retains the private IPv4 address, but a normal automatically assigned public IPv4 address can change. Check Elastic IP retention, association and release separately. Verify the actual destination instead of assuming the previous connection address still applies.

Charges for normally stopped compute are separate from EBS, snapshots, addresses, associated services and purchasing commitments. Termination does not automatically remove all of those either.

Choose an option to see the result and explanation.

After termination, can you start the same instance again to recover required data?
Read the answer and reason

Answer: No, establish required data handling before termination · A terminated instance cannot be restarted. Check each EBS volume’s retention/deletion setting and decide how required data is handled before terminating.

A short checklist before the action

  1. Choose the intended outcome: later restart, OS reboot, hibernation or permanent removal.
  2. Locate required data in EBS, instance store or RAM.
  3. Review each EBS termination-deletion setting and retention/recovery requirements.
  4. Check addresses, service dependencies, retained resources and commitments.
  5. Have an authorized operator act, then verify the state and remaining resources.

Check, then explain in your own words

  1. Explain which operation allows the same instance to be started again.
  2. Separate EBS, instance-store data and RAM.
  3. Review termination settings, addresses and retained charges.
Check your understanding: If an EBS-backed instance is stopped, does its instance-store data survive too?

No. Instance store is separate from retained EBS.

EBS can remain while a normal stop loses instance-store data and RAM. Establish the location of required data before choosing the operation.

Sources and conditions checked

The official references used in this lesson are collected here with their scope, conditions and corresponding sections. Independent expert review and live AWS comparison have not been performed.