NHN Cloud Associate flashcards
NHN Cloud Associate · Build recall with a focused review session.
32 cards. Read the term and the definition here. The deck at the top of this page is optional.
- Instance flavor
- A predefined combination of vCPU count and memory size used when launching an instance. The flavor family indicates whether the ratio favors general use, compute, or memory.
- Key pair
- An SSH public and private key pair injected into an instance at first boot. The private key stays with the user and cannot be recovered from the service if lost.
- Floating IP
- A separately allocated public IPv4 address that can be associated with or dissociated from an instance. Reassociation moves the address without recreating the instance.
- Security group
- A stateful virtual firewall applied to instances. It contains allow rules only; traffic matching no rule is dropped, and return traffic for allowed flows is permitted.
- Volume snapshot
- A point-in-time copy of a block storage volume. It can create a new volume in the same or another availability zone and is incremental after the first full copy.
- Auto scaling group
- A group of identical instances managed by a scaling policy. The policy sets a metric threshold and min/max size, and the group maintains the desired instance count.
- Container cluster
- A managed orchestration environment composed of a control plane and worker nodes. Workloads are scheduled as pods onto worker nodes that have sufficient free CPU and memory.
- Node pool
- A group of worker nodes sharing the same instance flavor and scaling configuration. Separate pools let workloads with different resource needs run in one cluster.
- Pod
- The smallest schedulable unit in a container cluster. Containers in the same pod share a network namespace and can share mounted volumes.
- Deployment
- A controller that maintains a desired replica count for stateless pods and performs rolling updates. It replaces failed pods automatically to keep the count.
- Service
- A stable virtual IP and DNS name in front of a changing set of pods. Clients address the service, not individual pod IPs, which change on restart.
- Config map and secret
- Objects that inject configuration into pods as environment variables or mounted files. Config maps hold non-sensitive values; secrets hold sensitive values such as passwords.
- Persistent volume claim
- A request for storage that binds to a persistent volume and outlives the pod using it. Deleting the pod does not delete the claim unless it is removed explicitly.
- Resource request and limit
- A request is the capacity the scheduler reserves for a pod; a limit is the runtime cap. A pod whose request exceeds free node capacity stays Pending.
- VPC
- The top-level isolated network boundary. Subnets inside it are each bound to one availability zone, so a multi-zone design needs subnets in multiple zones.
- Subnet CIDR
- The address range carved from the VPC block and assigned to one zone. The range cannot be resized after creation, so it must be planned for growth.
- VPC peering
- A direct connection between two VPCs that allows private routing between them. The two CIDR blocks must not overlap, or peering cannot be established.
- Route table
- A set of destination CIDR and next-hop pairs that decides where subnet traffic is sent. When several routes match, the most specific prefix is chosen.
- Network ACL
- A stateless filter applied at the subnet level. Because it keeps no connection state, return traffic must be allowed by an explicit rule in the opposite direction.
- Load balancer health check
- A periodic probe of backend instances on a defined protocol, port, and path. A backend failing the check is removed from rotation until it passes again.
- NAT gateway
- A managed egress point that lets instances without floating IPs start outbound internet connections. Inbound connections to those instances are not permitted through it.
- Block storage volume
- A disk attached to a single instance in the same availability zone. It can be detached and reattached elsewhere in the zone but never attached to two instances at once.
- Object storage bucket
- A container for objects addressed by key and accessed over an HTTP API rather than mounted as a disk. Access is controlled by bucket policy or signed URL.
- Bucket access policy
- The rule set governing who may read or write objects in a bucket. A public-read policy serves objects to anyone; a private bucket requires authentication or a signed URL.
- Lifecycle rule
- A bucket rule that transitions objects to colder storage or deletes them after a defined age. It is the standard control for log and backup retention cost.
- Volume performance
- For standard volume types, throughput and IOPS scale with allocated size. An undersized volume can bottleneck I/O even when the attached instance has spare capacity.
- Relational database instance
- A managed database provisioned with a flavor, storage size, and engine version. Storage can be expanded later, while engine upgrade options depend on the engine.
- Read replica
- A copy of a primary database that serves read queries. Replication is asynchronous, so the replica can lag and must not serve read-after-write consistent queries.
- Point-in-time recovery
- Restoring a database to a specific moment using automated backups. The recoverable moments are limited to the configured backup retention window.
- Database connection limit
- The maximum concurrent connections allowed by the instance flavor. Exceeding it returns connection errors even when CPU and memory utilization remain low.
- Database failover
- Promotion of a standby or replica when the primary fails. The endpoint changes, so the application must use retry logic and reconnect rather than reuse stale connections.
- Role-based authorization
- Access control where a role granted on a project or organization scope determines permitted actions. A missing role produces an authorization error, not a not-found response.