Storage, Data Protection, and Resiliency
Storage constructs and policies
This domain tests capacity and protection decisions. A storage pool is the set of physical disks; a container is a logical store with policies such as replication factor, compression, deduplication, and erasure coding. RF2 keeps two copies and tolerates one node failure. RF3 keeps three copies, tolerates two failures, and generally requires at least five nodes. EC-X saves capacity for cold data but adds compute overhead and is not ideal for hot, small random I/O. Snapshots are metadata-based and space-efficient because unchanged blocks are shared.
Data protection
Local snapshots are not backups. A Protection Domain groups VMs or vDisks for consistent snapshots and asynchronous replication to a remote cluster. RPO depends on the schedule. NearSync provides a lower RPO, measured in minutes, with bounded lag. Leap orchestrates DR runbooks. For ransomware or site loss, snapshots should be replicated or copied off-cluster.
Worked example
A 10 TB usable container with RF2 needs roughly 20 TB raw before overhead; with RF3 it needs roughly 30 TB. A 1 TB VM with 30 daily snapshots does not necessarily consume 30 TB, because unchanged blocks are shared. If 10% changes daily, incremental snapshot growth is about 100 GB per day, plus metadata.
Traps
Do not call a snapshot a backup. Do not deploy RF3 on a three-node cluster. Do not enable EC-X for active latency-sensitive workloads without testing. Do not assume deduplication and compression savings are identical for every workload; they are data-dependent.