MemberAugust 17, 2021 at 10:32 pm
“A high-performance computing (HPC) application has been launched in the company’s Amazon VPC. The application is composed of hundreds of private EC2 instances running in a cluster placement group, which allows the instances to communicate with each other at network speeds of up to 10 Gbps. There is also a custom cluster controller EC2 instance that closely controls and monitors the system performance of each instance. The cluster controller has the same instance type and AMI as the other instances. It is configured with a public IP address and running outside the placement group. The Solutions Architect has been tasked to improve the network performance between the controller instance and the EC2 instances in the placement group.
Which option provides the MOST suitable solution that the Architect must implement to satisfy the requirement while maintaining low-latency network performance?”
As per the practice test, the answer to above question is to “Stop the custom cluster controller instance and move it to the existing placement group”. However, as per the AWS documentation https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/placement-groups.html
“If you try to add more instances to the placement group later, or if you try to launch more than one instance type in the placement group, you increase your chances of getting an insufficient capacity error. If you receive a capacity error when launching an instance in a placement group that already has running instances, stop and start all of the instances in the placement group, and try the launch again. Starting the instances may migrate them to hardware that has capacity for all of the requested instances.”
Due to this, shouldn’t the right answer be “<strong style=”font-family: inherit; font-size: inherit;”>Terminate the custom cluster controller EC2 instance and stop all of the running instances in the existing placement group. Move the cluster controller instance to the existing placement group and restart all of the instances”
MemberAugust 18, 2021 at 9:47 pm
Thanks for the feedback.
You are correct that if you want to add an instance to a placement group, it is recommended to use a single launch request to launch all the number of instances you need.
However, this is a custom controller EC2 which may not be the OS setup (or installed applications) as the other instances inside the Placement-group. You will want to lessen your work and don’t want to recreate that custom controller, and possibly spend more time troubleshooting after recreating that instance. So you can first try to add it manually to the existing placement group.
Additionally, AWS now allows you to move an EC2 to an existing placement group https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/placement-groups.html#change-instance-placement-group.
(This is the documentation where used as reference for creating this question scenario)
The chance of getting an insufficient capacity error is not fixed. If it is just one instance, there is a good chance that the current placement group host can accommodate another instance.
Hope this helps.
Kenneth Samonte @ Tutorials Dojo
Log in to reply.