Protection Sources and Policy Configuration
Protection Sources and Policy Configuration
This domain tests how you register workloads and translate business requirements into a protection policy and group. Exam items usually give you an SLA statement and ask which policy settings satisfy it, or show a source hierarchy and ask what a group will actually protect.
Sources and registration
Supported protection sources include VMware vSphere (via VADP with changed block tracking), Hyper-V, physical Windows and Linux servers (Cohesity agent), NAS shares over SMB/NFS, Microsoft 365, enterprise databases such as SQL Server, Oracle, and SAP HANA, cloud workloads, and Kubernetes. Registering a source requires credentials and connectivity; the source tree is hierarchical, so registering a vCenter exposes datacentres, clusters, hosts, and VMs, and a group can select any node in that tree.
Policy anatomy
A protection policy is SLA-based. Its settings are:
- RPO / frequency - how often snapshots are taken (for example every 4 hours).
- Retention - how long local snapshots are kept.
- Replication - target cluster and retention there.
- Archival - external target and retention (extended retention lives here).
- Full backup schedule - when synthetic or full baselines are produced.
- Indexing - enables fast file-level search and restore.
- Exclusions and QoS - excluded disks or paths, throughput limits.
A protection group binds selected sources to one policy, one storage domain, and run options. Policy changes are read at the start of each run, so they affect future snapshots, not existing ones.
Worked example
Requirement: "No more than four hours of data loss; keep 30 days locally; keep monthlies for seven years; keep a copy at the DR site." Reasoning: 4 hours sets RPO to a 4-hourly schedule. 30 days sets local retention. Seven years of monthlies cannot be local retention - it maps to archival with a monthly schedule and a 7-year archive retention. The DR copy maps to replication to the second cluster. One policy expresses all four requirements.
Traps that cost marks
Treating RPO and retention as the same setting; assuming indexing is on by default; expecting a policy edit to rewrite already-captured snapshots; forgetting that extended retention requires an archive target; deleting a policy that groups still reference.