Huawei HCIP-Datacom common mistakes
Huawei HCIP-Datacom · Understand the traps and how to avoid them.
Treating OSPF neighbor adjacency as proof that routing works, then failing questions where routers are adjacent but no OSPF route appears in the routing table because the network was advertised under the wrong area or the interface was passive.
Candidates memorize the seven neighbor states and stop there, so they never drill the conditions that actually install a route: area match, matching network command, non-passive interface, and route preference versus other protocols.
What to do instead
For each OSPF scenario, walk a fixed checklist: neighbor state, LSDB contents, area ID on both ends, network statement coverage, passive-interface status, and route preference against static and IS-IS routes. Practice reading display ospf peer, display ospf lsdb, and display ip routing-table protocol ospf output together rather than in isolation.
Choosing the wrong DR/BDR outcome on broadcast segments, for example assuming the highest router ID always wins regardless of priority, or forgetting that priority 0 makes a router ineligible and that preemption does not occur after a stable election.
The election rules are simple but conditional, and candidates learn a single slogan instead of the full order: priority first, then router ID, with no preemption and a waiting timer before election.
What to do instead
Rehearse the exact sequence: exclude priority-0 routers, compare priority, break ties by router ID, and remember the 40-second wait timer and absence of preemption. Work through three-router examples where one router reboots and the election does not change.
Confusing IS-IS Level-1 and Level-2 behavior, such as expecting a Level-1 router to learn inter-area routes directly or misidentifying which router sets the ATT bit and why.
IS-IS terminology (NET, LSP, DIS, ATT bit, default route injection) differs from OSPF, and candidates map OSPF concepts onto IS-IS one-to-one, which breaks on area boundaries and default route propagation.
What to do instead
Build a two-level topology and trace how a Level-1-only router reaches another area: L1/L2 router sets ATT bit in its L1 LSP, L1 routers install a default route toward it. Verify with display isis lsdb and display isis route, and note that IS-IS has no backbone requirement equivalent to OSPF area 0 in the same sense.
Mixing up BGP path-selection order, especially placing local preference after AS-path length or forgetting that weight is Cisco-style and not the first criterion on Huawei equipment, where preferred value plays that role.
The selection sequence is long and vendor-flavored; candidates memorize a list from another vendor or from an older document and apply it unchanged.
What to do instead
Memorize the Huawei order explicitly: preferred value, local preference, locally originated, AS-path length, origin type, MED, EBGP over IBGP, lowest IGP metric to next hop, oldest route, lowest router ID, shortest cluster list, lowest peer address. Then test yourself with scenarios where two attributes conflict and only one tiebreaker decides.
Answering BGP reachability questions without checking the next-hop reachability rule, for example expecting an IBGP-learned route to be best when the next hop is unreachable, or forgetting that EBGP next hop is unchanged when advertised to IBGP peers unless next-hop-local is configured.
Candidates focus on attributes and ignore that BGP will not use a route whose next hop cannot be resolved in the IGP, and they forget the IBGP next-hop preservation rule.
What to do instead
Whenever a BGP route is not selected, check display bgp routing-table for the route's validity flag and resolve the next hop in the IGP table. Practice the standard fix: peer x.x.x.x next-hop-local on the route reflector or AS boundary router, plus an IGP route to the external next hop.
Configuring an Eth-Trunk in manual load-balance mode and expecting LACP negotiation, or mismatching LACP modes (active/passive) and wondering why the trunk stays down.
Huawei offers manual and LACP modes with several LACP variants, and candidates assume any two trunk ends will bundle because the physical links are up.
What to do instead
Learn the matrix: manual mode requires both ends manual; LACP active works with active or passive; passive with passive never negotiates. Also confirm member interfaces match in speed, duplex, and VLAN configuration, and check display eth-trunk for member status before blaming the peer.
Treating RSTP and MSTP as interchangeable, so questions about edge ports, P/A negotiation, or instance-to-VLAN mapping are answered with STP timers and port states that do not apply.
Candidates learn classic STP first and carry its blocking/listening/learning model and 30-second convergence into RSTP and MSTP questions.
What to do instead
Separate the three: RSTP has discarding/learning/forwarding and proposal-agreement for fast convergence; MSTP adds regions, instances, and the IST. Practice mapping VLANs to instances with instance 1 vlan 10 and verifying with display stp instance 1 and display stp region-configuration.
Misconfiguring VLAN assignment on access versus trunk ports, then misreading hybrid port behavior, which leads to wrong answers about which frames are tagged on egress.
Huawei hybrid ports use tagged and untagged VLAN lists that behave differently from trunk and access, and candidates apply switchport-style reasoning from other vendors.
What to do instead
Write out the three port types side by side: access sends untagged for one VLAN; trunk sends tagged for allowed VLANs with one untagged PVID; hybrid sends tagged or untagged per VLAN list. Then trace a frame in each direction across a hybrid link and confirm with display port vlan.
Assuming VXLAN tunnels come up as soon as VTEPs are configured, without verifying that the NVE peers are reachable over the underlay and that the VNI is mapped to a valid VNI-to-VLAN binding.
Candidates learn the VXLAN header format but skip the operational dependencies: underlay routing to the remote VTEP, source interface selection, and the service access point binding.
What to do instead
For any VXLAN scenario, check in order: underlay route to the remote VTEP address, source interface on the NVE, VNI configured, and bridge-domain to VNI binding. Use display vxlan vni and display nve peers to confirm peer state before analyzing encapsulation.
Answering route-policy and ACL questions by reading the first matching rule only, ignoring implicit deny, default permit behavior differences, and the fact that a route-policy with no matching node denies by default while an empty ACL permits.
Filtering tools have asymmetric defaults, and candidates memorize syntax without the deny/permit semantics of unmatched traffic.
What to do instead
Build a comparison table: ACL unmatched traffic is permitted; route-policy unmatched routes are denied; a node with no permit/deny statement but only if-match conditions permits when matched. Practice writing a route-policy that filters specific prefixes and permits everything else, then verify with display route-policy and route counters.