Compute
This section covers NHN Cloud Compute: how instances are created, sized, accessed, scaled, and imaged. The exam checks whether you can choose the right compute resource and reason about lifecycle and networking dependencies.
Core rules
- An instance is launched from an image and a flavor. The flavor fixes vCPU and memory; the image supplies OS and basic software. Changing flavor usually requires stopping the instance, so plan capacity before launch.
- Key pairs are injected at creation for SSH access. Losing the private key does not let you recover it from NHN Cloud; create a new key pair and rebuild access.
- Instances live in a VPC subnet and receive a private IP. A public IP or floating IP is separate; attaching one does not change the private address.
- Instance templates and Auto Scaling groups let you repeat launches and scale by policy. The template captures image, flavor, key pair, subnet, and security groups.
- Snapshots/images preserve boot data; block volumes are separate and must be backed up according to their own policy.
Worked example
A web tier needs three identical Linux instances that scale out when CPU is high. Create an instance template with the approved image, flavor, key pair, private subnet, and web security group. Create an Auto Scaling group from the template with min 2, desired 3, max 6, and a CPU target policy. Place a load balancer in front. Reason: the template removes configuration drift; the group supplies identical members; the load balancer health check removes failed nodes.
Traps
Do not confuse a public IP with a private IP. Do not assume a flavor change is online. Do not put a database on a public subnet just to reach it. Do not forget that an Auto Scaling group needs a healthy load-balancer target group to route traffic.