10% of the CCA exam. Sample questions below; the full library has 19 questions tagged to this domain.
Sample questions on Installation and Configuration
Installation and Configuration
Q1. You run 'cilium connectivity test' and one of the test pods is in CrashLoopBackOff state. The logs show 'BPF verification failed'. What is the likely issue?
Reveal answer and explanations
AThe kernel version is too old and doesn't support the eBPF programs Cilium is trying to load
Correct. BPF verification failures typically indicate kernel incompatibility or missing eBPF support features in the running kernel version.
BThe connectivity test requires a specific kernel capability (CAP_SYS_ADMIN) that wasn't granted
Incorrect. Capability issues would cause permission errors, not BPF verification failures.
CThe test pod's network namespace wasn't properly isolated from the host namespace
Incorrect. Namespace isolation issues would cause network connectivity failures, not BPF verification errors.
DThe cilium-agent failed to attach the BPF program to the test pod's veth interface
Incorrect. BPF program attachment errors are logged separately from verification failures.
Installation and Configuration
Q2. You want to enable Cilium's DNS policy enforcement after installation. Which Helm value must be set, and what is the prerequisite?
Reveal answer and explanations
ASet 'l7Proxy=true' and 'dnsPolicyEnforcement=true'; prerequisite is running with kubeProxyReplacement
Correct. DNS policy enforcement requires L7 proxying (l7Proxy=true) and DNS policy enforcement (dnsPolicyEnforcement=true), with kubeProxyReplacement recommended.
BSet 'l7Protocols=dns'; no prerequisite, but it only works with the Cilium ingress controller
Incorrect. DNS policies are independent of the ingress controller.
CSet 'dnsPolicy=true'; prerequisite is DNS caching enabled in kube-dns
Incorrect. DNS policy uses L7 proxying, not a simple boolean flag.
DSet 'dnsSecurity=true'; prerequisite is all pods using CoreDNS (not kube-dns)
Incorrect. CoreDNS vs kube-dns doesn't matter for DNS policy enforcement; it works with either.
Installation and Configuration
Q3. After installing Cilium, you run 'cilium endpoint list' but see no endpoints listed, even though pods are running. What should you check first?
Reveal answer and explanations
AVerify that cilium-agents are running on all nodes
Correct. If cilium-agents aren't running, endpoints won't be discovered or managed; this is the first check.
BEnsure the Cilium operator is running and has synced with the API server
Incorrect. Operator issues wouldn't prevent agent endpoint discovery and listing.
CConfirm that pods have been annotated with 'cilium.io/inject=true'
Incorrect. Pod injection annotations aren't required for Cilium to manage endpoints.
DCheck if the kubeconfig is correctly pointing to the cluster
Incorrect. kubeconfig issues would prevent any cluster access, not just endpoint listing.
Installation and Configuration
Q4. You want to change the Cilium datapath tunnel type from VXLAN to Geneve after installation. Which command(s) would you use?
Reveal answer and explanations
A'cilium-dbg config set tunnel geneve' followed by agent restart
Incorrect. Tunnel protocol can be changed without a full reinstall.
CUpdate the cilium-config ConfigMap and trigger a rolling restart of cilium-agents
Correct. The cilium-config ConfigMap is the persistent configuration source; changing tunnelProtocol there and restarting agents applies the change cluster-wide.
D'cilium config set tunnel geneve' on each node
Incorrect. Local cilium config command doesn't persist changes across agent restarts.
Installation and Configuration
Q5. You run 'cilium install --set kubeProxyReplacement=true --version 1.15.1'. The installation hangs during the operator initialization phase. What is the most common root cause?
Reveal answer and explanations
AThe cilium Helm chart version 1.15.1 doesn't support kubeProxyReplacement
Incorrect. Cilium 1.15.1 supports kubeProxyReplacement; this isn't a version compatibility issue.
BThe kube-proxy DaemonSet is still running and conflicts with kubeProxyReplacement initialization
Correct. When kube-proxy is still running, kubeProxyReplacement initialization hangs due to port conflicts; kube-proxy must be scaled down or removed first.
CThe kernel doesn't support the eBPF features required by kubeProxyReplacement in this version
Incorrect. Kernel eBPF feature checks happen during agent initialization, not operator initialization.
DThe API server is unreachable; the operator can't communicate with the cluster
Incorrect. Operator would fail fast with API server connectivity issues, not hang.
Installation and Configuration accounts for 10% of the CCA exam. Expect questions that test recall of terminology and the ability to read short scenarios — not deep configuration. Use the sample questions above as difficulty calibration; if any feel hard, the rest of our 19-question domain bank will close those gaps.