Exam AWS DevOps Engineer Professional topic 1 question 94 discussion

https://repost.aws/knowledge-center/auto-scaling-group-rolling-updates Option B is incorrect because the AutoScalingReplacingUpdate policy terminates all instances in the Auto Scaling group simultaneously and creates new instances, which does not satisfy the requirement of having at least four application servers running at all times.

its A like for Apple.

A is correct Autoscaling Rolling Update: Autoscaling Rolling Update is a mechanism that allows for updates to be applied to instances in an autoscaling group in a rolling fashion. This means that a few instances are taken down at a time, updated, and then brought back up before the next set of instances are taken down. This process continues until all instances have been updated. Autoscaling Rolling Update helps to ensure that there is no downtime and that the application remains available throughout the update process. Autoscaling Replacing Update: Autoscaling Replacing Update is a mechanism that replaces the entire set of instances in an autoscaling group with new instances. Unlike Autoscaling Rolling Update, this method takes down all the old instances at once and replaces them with new instances. Autoscaling Replacing Update is faster than Autoscaling Rolling Update since it replaces all instances at once, but there is a period of downtime during the update process.

Only B satisfies the scenario: "After the stack update finishes, the engineer manually terminates the old instances one by one, verifying that the new instance is operational before proceeding. " https://docs.aws.amazon.com/AWSCloudFormation/latest/UserGuide/aws-attribute-updatepolicy.html#cfn-attributes-updatepolicy-replacingupdate

Both A and B cannot guarantee that "at least four application servers running at all times". C is involved by D.

B - correct (in my opinion) - requirement: LEAST number of manual steps moving forward Solution: use AutoScalingReplacingUpdate policy and set "AutoScalingReplacingUpdate" : { "WillReplace" : "true" } The AutoScalingReplacingUpdate will be terminating instances ONLY after new instances are available and stable. A - also correct to some extent but the potential problem is related to roll back - with the rolling update you have to take care of auto scaling group configuration Note: If an unexpected scaling action changes the state of the Auto Scaling group during a rolling update, the update can fail. The failure can result from an inconsistent view of the Auto Scaling group by AWS CloudFormation.

A is correct because of the ability to specify MinInstanceInService attributes on the autoscalinggroup update policy.

AAAAAAAAAAAAAA

For rolling updates, you can specify whether AWS CloudFormation updates the instances in the Auto Scaling group in batches, or all instances at once.