CCIE Data Center cheat sheet

Trace VLAN, gateway, VRF, route, next hop, policy, and return path separately.

vPC peer-link and keepalive serve different roles; inspect actual peer and consistency state.

Carrier up does not prove VLAN continuity or port-channel member compatibility.

Per-flow hashing does not promise one flow the entire bundle rate.

OSPF state must be interpreted with network type, roles, timers, area, and MTU.

BGP established does not prove accepted routes or reachable next hops.

Longest-prefix lookup is distinct from protocol preference for the same prefix.

Multicast requires appropriate membership, tree, rendezvous, and RPF state.

Tenant, VRF, bridge domain, EPG, and attachment are related but distinct objects.

Trace interface selector, policy group, AEP, domain, pool, and EPG binding.

Compare actual server tags and virtual attachments with intended classification.

Routes do not automatically authorize inter-EPG services in an enforced policy model.

Check contract direction relationships and actual protocol/port filters.

Same-EPG behavior has its own isolation and contract exceptions.

APIC is management and policy control, not the inline hop for all packets.

Correlate faults, events, health, audit, and endpoint state.

Underlay reachability and MTU must support the actual VXLAN transport.

EVPN type 2 advertises MAC/IP, type 3 membership, and type 5 IP-prefix information.

RD uniqueness and RT membership policy have different roles.

Check L2 VNI mapping separately from VRF/L3 VNI routed behavior.

A live peer does not prove every VNI is operational.

ARP suppression and mobility require trustworthy bindings and location state.

External routes and external EPG contracts are separate checks.

Multi-Pod, Multi-Site, and remote leaf have different boundaries and requirements.

Profiles represent desired identities and policies; verify deployment and activation.

Initial and updating templates differ in later change propagation.

Pool capacity and actual identity allocation must match profile requirements.

Inspect maintenance acknowledgment and reboot implications before policy activation.

Trace Ethernet VLANs and FC VSANs as distinct service paths.

Validate SAN boot identity, target, LUN, boot mode, and OS behavior.

Host teaming and multipathing require appropriate host configuration.

Compatibility and independent failure domains matter more than version-number or cable counts.

Adapter visibility, fabric login, active zoning, and LUN presentation are separate.

Configured zones do not imply membership in the active zoneset.

A replacement pWWN needs deliberate access-policy reconciliation.

NPV relies on suitable upstream NPIV behavior.

VSAN trunk membership matters even when the physical ISL is up.

FC credits and distance can limit sustained throughput.

RoCEv2 routing does not remove QoS and congestion-control requirements.

PFC pause and ECN-driven source adaptation solve different parts of congestion handling.

Authentication and authorization are distinct; validate actual role mapping.

ACL order and complete forward/return paths affect service permission.

First-hop trust and binding sources must cover legitimate hosts correctly.

Mirroring direction and transport influence capture interpretation.

Long utilization averages can hide microbursts and queue drops.

Policing, shaping, scheduling, marking, and CoPP have different purposes.

Stateful service insertion needs the intended bidirectional behavior.

Protect tokens, cookies, keys, and sensitive operational exports.

Validate target environment, object identity, types, and supported values before writes.

HTTP success is not enough; inspect command or managed-object results.

Handle auth expiry, pagination, cardinality, and asynchronous task completion.

Respect rate feedback and bounded retry budgets.

Reconcile uncertain write outcomes before repeating non-idempotent operations.

Preview coverage depends on module support and does not guarantee nondisruption.

Preserve state, dependencies, per-target outcomes, and tested recovery procedures.

Verify active configuration and relevant application behavior after the approved change.