Manage Amazon EC2 instances scheduled for reboot
When AWS must perform tasks such as installing updates or maintaining the underlying host, it can schedule an instance reboot. During the scheduled reboot, the instance either stays on the same host, or migrates to a different host, depending on the event, as follows:
-
instance-rebootevent-
During the reboot, the instance remains on the host. This is known as an in-place reboot.
-
The current host undergoes maintenance.
-
Typically completes in seconds.
-
-
system-rebootevent-
During the reboot, the instance is migrated to a new host. This is known as a reboot migration.
-
Typically completes in minutes.
-
To check what type of event is scheduled for your instance, see Determine the event type.
Actions you can take
When you receive a scheduled instance-reboot or
system-reboot event notification, you can take one of the
following actions:
-
Wait for scheduled reboot: You can wait for the instance reboot to occur within its scheduled maintenance window.
-
Reschedule the reboot: You can reschedule the instance reboot to a date and time that suits you.
-
Perform a user-initiated reboot: You can manually reboot the instance yourself at a time that suits you. However, the outcome depends on the event:
-
instance-rebootevent – Your instance remains on the current hardware (in-place reboot), no host maintenance takes place, and the event stays open. -
system-rebootevent-
If reboot migration is enabled on your instance, a user-initiated reboot attempts to migrate your instance to new hardware. If successful, the event is cleared. If unsuccessful, an in-place reboot occurs and the event remains scheduled.
-
If reboot migration is disabled on your instance, a user-initiated reboot keeps the instance on the same hardware (in-place reboot), no host maintenance takes place, and the event remains scheduled. When the scheduled event eventually takes place, AWS will move your instance to new hardware (reboot migration).
-
-
After AWS reboots your instance
The following applies after AWS reboots your instance:
-
The scheduled event is cleared.
-
The event description is updated.
-
For an
instance-rebootevent:-
Maintenance of the underlying host is complete.
-
-
For a
system-rebootevent:-
The instance moves to a new host.
-
The instance retains its IP address and DNS name.
-
Any data on local instance store volumes is preserved.
-
-
You can use your instance after it has fully booted.
Alternative options
If you can't reschedule the reboot event or enable reboot migration for a user-initiated reboot, but want to maintain normal operation during the scheduled maintenance window, you can do the following:
-
For an instance with an EBS root volume
-
Manually stop and start the instance to migrate it to a new host. This is not the same as manually rebooting the instance, where the instance stays on the same host.
-
Optionally, automate an immediate instance stop and start in response to the scheduled reboot event. For more information, see Running operations on EC2 instances automatically in response to events in AWS Health in the AWS Health User Guide.
Important
The data on instance store volumes is lost when an instance is stopped. For more information, see Stop and start Amazon EC2 instances.
-
-
For an instance with an instance store root volume
-
Launch a replacement instance from your most recent AMI.
-
Migrate all necessary data to the replacement instance before the scheduled maintenance window.
-
Terminate the original instance.
-
Enable or disable reboot migration
When an instance is scheduled for a system-reboot event, you can
reboot it before the event. The outcome of a user-initiated reboot depends on
the instance's reboot migration setting:
-
Enabled – A user-initiated reboot attempts to migrate your instance to new hardware (reboot migration). If successful, the event is cleared. If unsuccessful, an in-place reboot occurs and the event remains scheduled. Note that even when enabled, reboot migration can only occur if your instance meets the reboot migration requirements.
-
Disabled – A user-initiated reboot keeps the instance on the same hardware (in-place reboot), no host maintenance takes place, and the event remains scheduled. When the scheduled event eventually takes place, AWS will move your instance to new hardware (reboot migration).
A reboot with migration takes longer than an in-place reboot:
-
In-place reboot: Approximately 30 seconds
-
Reboot with migration: Several minutes
Note
Instances that receive a system-reboot event notification are
enabled for user-initiated reboot migration by default.
Requirements for enabling reboot migration
Reboot migration can be enabled on instances that meet the following criteria:
- Instance types
-
Not all instance types support enabling reboot migration. You can view the instance types that support enabling reboot migration.
- Tenancy
-
-
Shared
-
Dedicated Instance
For more information, see Amazon EC2 Dedicated Instances.
-
Limitations
Reboot migration is not supported for instances with the following characteristics:
-
Platform: Instances running natively on the Xen hypervisor
-
Instance size:
metalinstances -
Tenancy: Dedicated Host. For Dedicated Hosts, use Dedicated Host Auto Recovery instead.
-
Storage: Instances with instance store volumes
-
Networking: Instances using an Elastic Fabric Adapter
-
Auto Scaling: Instances that are part of an Auto Scaling group
Steps for enabling or disabling reboot migration
When an instance receives a system-reboot event, it is enabled
for reboot migration by default. You can disable reboot migration so that
during a user-initiated reboot, the instance stays on the same hardware
(in-place reboot).
The default configuration doesn't enable reboot migration for an
unsupported instance. For more information, see Requirements for enabling
reboot migration.
You can disable or enable reboot migration on a running or stopped instance.
