# Free VO practice questions

> 10 free VO practice questions with the correct answer and an explanation.

Page: https://devopscodelab.com/exams/vault-operations/free-practice-questions/
Practice exams: https://devopscodelab.com/exams/vault-operations/

### Question 1 of 10 (choose 2)

Following production hardening guidance, which two changes should be applied to a Vault server that uses integrated storage on Linux? (Choose two.)

- A. Disable swap on the host and set disable_mlock = true, since memory locking is not recommended with integrated storage
- B. Enable raw_storage_endpoint so operators can repair storage entries directly when a mount is corrupted
- C. Keep the initial root token in a sealed envelope so it can be reused for routine administrative work
- D. Run the Vault process as a dedicated unprivileged user and disable core dumps for that process
- E. Terminate TLS at the load balancer and use plain HTTP listeners on the nodes to reduce certificate sprawl

**Answer:** A, D

**Explanation:** Hardening guidance says to run Vault as an unprivileged user, prevent core dumps that could expose key material, and, with integrated storage, disable swap at the OS level while setting disable_mlock because memory locking interacts badly with the memory-mapped raft database. The raw storage endpoint should stay disabled because it bypasses access controls. TLS should be end to end rather than terminated before the nodes. Root tokens should be revoked after setup and regenerated only for emergencies.

### Question 2 of 10

A team configures transit auto-unseal so that a central Vault cluster unseals several regional clusters. Regional clusters must keep unsealing after restarts, even when they restart weeks later. Which token should the regional seal stanza use?

- A. A periodic, orphan token limited by a policy to the encrypt and decrypt endpoints of one transit key
- B. A batch token with a long TTL limited by a policy to the encrypt and decrypt endpoints of one transit key
- C. A non-periodic child token of an operator's personal token with a 32-day explicit maximum TTL set
- D. A service token created by the regional cluster's own root token and given the default policy only

**Answer:** A

**Explanation:** Transit seal tokens must stay valid indefinitely, so a periodic orphan token that Vault renews is recommended, and a tightly scoped policy limits blast radius. Batch tokens cannot be renewed and expire at their TTL, which eventually breaks unsealing. A token from the regional cluster cannot authenticate to the central cluster, and the default policy does not allow transit operations. A child of a personal token dies when that token is revoked and has a hard maximum lifetime.

### Question 3 of 10

A Prometheus server must scrape Vault metrics without holding a Vault token. Which configuration enables this?

- A. Set prometheus_retention_time in the telemetry stanza and enable the statsd_address option so Prometheus can pull without auth
- B. Set disable_hostname in the telemetry stanza and grant the anonymous policy read on sys/metrics through an ACL
- C. Set prometheus_retention_time in telemetry and unauthenticated_metrics_access = true in the listener telemetry block
- D. Set unauthenticated_metrics_access in the seal stanza and expose sys/metrics on the cluster port 8201 instead

**Answer:** C

**Explanation:** Prometheus format metrics require prometheus_retention_time to be non-zero, and the listener telemetry block's unauthenticated_metrics_access lets sys/metrics be scraped without a token. statsd_address pushes metrics to a statsd agent and does not affect authentication on sys/metrics. There is no anonymous policy that applies to unauthenticated requests, and disable_hostname only changes metric names. The seal stanza and the cluster port are not used to expose metrics.

### Question 4 of 10

A team runs Vault itself on Kubernetes with the official Helm chart. Which practice MOST reduces the risk that another workload compromises Vault's memory or storage?

- A. Run Vault pods with hostNetwork enabled so traffic avoids the overlay network
- B. Run Vault as a privileged container so that mlock can lock pages into memory
- C. Store the unseal keys in a ConfigMap so pods can restart without manual action
- D. Schedule Vault pods on dedicated nodes using taints, tolerations and affinity

**Answer:** D

**Explanation:** Isolating Vault on dedicated nodes prevents co-located workloads from accessing the same host, which is the main multi-tenancy risk for Vault on Kubernetes. hostNetwork exposes host networking without improving isolation. ConfigMaps are plain text and readable by anyone with namespace access, which is far worse than auto-unseal. Privileged mode gives far more than the IPC_LOCK capability that mlock needs, increasing the attack surface.

### Question 5 of 10

After a DR failover, the promoted cluster is primary and the original primary has been demoted. The team now wants the demoted cluster to follow the new primary as its DR secondary. What is the correct procedure?

- A. Generate a secondary activation token on the new primary and call update-primary on the demoted cluster with it
- B. Restore a raft snapshot from the new primary onto the demoted cluster and enable DR primary mode there afterwards
- C. Run sys/replication/dr/secondary/enable on the demoted cluster with its original root token as the activation token
- D. Promote the demoted cluster again and call update-primary on the new primary so the roles swap back automatically

**Answer:** A

**Explanation:** A demoted primary becomes a DR secondary that can be pointed at the new primary through sys/replication/dr/secondary/update-primary, using a secondary activation token generated on the new primary. Enabling replication fresh would wipe and resync it and requires a token from the primary rather than a root token. Restoring a snapshot does not set up replication. Promoting it again would create two primaries.

### Question 6 of 10

A Vault Enterprise primary cluster in us-east-1 becomes permanently unavailable. A DR secondary in us-west-2 is healthy. The team created a batch DR operation token in advance. Which steps restore service with the LEAST delay?

- A. Promote the DR secondary using a regular admin token from the primary, then copy the primary's leases manually
- B. Promote the DR secondary with the batch DR operation token, then repoint clients or DNS to it
- C. Enable performance replication on the DR secondary, then let clients authenticate there with new credentials
- D. Generate a DR operation token on the secondary with recovery key holders, then demote the failed primary first

**Answer:** B

**Explanation:** A batch DR operation token prepared ahead of time can authorize sys/replication/dr/secondary/promote immediately, without assembling key holders, and DNS or the load balancer then sends clients to the new primary. Generating an operation token with key holders works but takes longer, and the failed primary cannot be demoted while unreachable. DR secondaries do not accept normal client tokens for promotion, and leases are already replicated. Performance replication does not make a DR secondary serve requests.

### Question 7 of 10

A financial services firm wants Vault Enterprise to auto-unseal using its existing on-premises HSM and also to protect PKI CA private keys with the HSM boundary for FIPS 140-3 alignment. Which feature combination meets both goals?

- A. A pkcs11 seal stanza for auto-unseal plus response wrapping, which stores PKI keys inside the HSM's key slots
- B. A pkcs11 seal for auto-unseal plus seal wrap, which adds HSM-backed encryption around sensitive values
- C. A transit seal stanza for auto-unseal plus entropy augmentation, which generates PKI keys directly inside the HSM
- D. A Shamir seal with HSM-held unseal shares plus the HSM secrets engine, which moves PKI CA keys onto the HSM

**Answer:** B

**Explanation:** The PKCS#11 seal lets the HSM protect the root key for auto-unseal, and seal wrap adds an extra encryption layer from the seal around sensitive values such as PKI CA keys, supporting FIPS-aligned deployments. A transit seal uses another Vault cluster, not the HSM, and entropy augmentation only supplies randomness. Response wrapping protects responses in transit and does not involve the HSM. Shamir shares are not held by an HSM, and there is no HSM secrets engine for this purpose.

### Question 8 of 10

A read-heavy application hammers a Vault Enterprise cluster with KV reads, and the active node's CPU is saturated while the two standbys idle. Which feature lets the standbys serve those reads?

- A. A rate limit quota on the KV mount so the active node is relieved
- B. Batch tokens for the application, so reads skip token lookups
- C. Performance standby nodes, which handle read requests locally
- D. Performance replication to a secondary cluster in the same region

**Answer:** C

**Explanation:** Performance standbys are Enterprise standby nodes that service requests that do not modify storage, spreading read load across the cluster. A performance secondary would require clients to target another cluster. Batch tokens reduce token storage writes but reads are still processed on the active node. A rate limit quota rejects excess requests rather than adding capacity.

### Question 9 of 10

In Vault Enterprise, the team wants every write to kv/prod/* to be allowed only when the requesting entity belongs to the prod-deployers group, regardless of ACL grants. Sentinel is available. Which policy type fits best?

- A. A control group on kv/prod/* requiring one approval from a prod-deployers member
- B. An RGP attached only to the prod-deployers group that grants write on kv/prod/*
- C. An EGP on kv/prod/* checking identity.groups for prod-deployers
- D. An ACL deny on kv/prod/* for non-deployers

**Answer:** C

**Explanation:** An endpoint governing policy applies to every request on the path and can check the requester's group membership, rejecting anyone outside prod-deployers. ACL deny cannot express a group condition. An RGP attached to the group only affects its members and cannot restrict others. A control group adds approvals rather than restricting who may write.

### Question 10 of 10

A team runs Vault Agent with a template that reloads nginx after rendering certificates. Nginx sometimes fails to reload because the command runs before both files exist. Which approach resolves this?

- A. Render certificate and key in one file, or run the reload only after the last template
- B. Set the sink mode to 0600 so nginx cannot read partially written files during a reload of the service
- C. Set exit_after_auth = true so templates render once and nginx is reloaded by systemd
- D. Set static_secret_render_interval to a higher value so files are rendered more slowly

**Answer:** A

**Explanation:** Commands run per template, so rendering both values in one file or triggering the reload only after the final file avoids reloading with incomplete data. Exiting after auth stops renewals. The static render interval does not apply to PKI certificates fetched with pkiCert and does not order renders. Sink mode affects token files, not template output.

Ready for a timed run? Take the free 15-question VO mini exam: https://devopscodelab.com/app/exams/vault-operations/mini/

DevOpsCode Lab is not affiliated with HashiCorp. VO and related names are trademarks of their owners. We write every question ourselves; these are not real exam questions.
