Review advertised-route output, prefix-list matches and outbound route-policy order. This check would diagnose BGP prefix export-filter leak, but would not first isolate the described PBR failed-next-hop fallback. Begin with: Check PBR next hop, track state, route resolution and tested failure behavior.

Inspect PBR classifier, interface direction, policy hits and a traffic trace. This check would diagnose Policy-based routing for approved application, but would not first isolate the described PBR failed-next-hop fallback. Begin with: Check PBR next hop, track state, route resolution and tested failure behavior.

Inspect route-policy node IDs, if-match criteria, apply clauses and hit counts. This check would diagnose Route-policy permit and deny node order, but would not first isolate the described PBR failed-next-hop fallback. Begin with: Check PBR next hop, track state, route resolution and tested failure behavior.

Apply supported route-policy attributes only to the documented matched prefixes and validate next-hop resolution. This is a legitimate response to the different issue of Route-policy next-hop and preference application, but it does not correct the specific MQC classifier and behavior association condition reported. The relevant technical distinction is: Huawei MQC separates class identification, traffic behavior and applied traffic-policy binding.

Apply supported next-hop handling such as next-hop-local where appropriate and provide underlay reachability. This is a legitimate response to the different issue of iBGP next-hop reachability, but it does not correct the specific MQC classifier and behavior association condition reported. The relevant technical distinction is: Huawei MQC separates class identification, traffic behavior and applied traffic-policy binding.

Align the approved destination NAT mapping and corresponding policy with the intended public service. This is a legitimate response to the different issue of Destination NAT and security-policy pair, but it does not correct the specific MQC classifier and behavior association condition reported. The relevant technical distinction is: Huawei MQC separates class identification, traffic behavior and applied traffic-policy binding.

Every prefix has a stable authoritative origin and route import does not recursively reappear. That observation would support a separate case of Redistribution route tag loop avoidance, not verification of the reported MQC classifier and behavior association. The relevant observable outcome here is: Matched voice flows receive the configured treatment and unrelated traffic remains within its approved class.

Clients obtain a valid address lease, gateway and DNS options from the authorized remote DHCP server. That observation would support a separate case of DHCP relay to remote server, not verification of the reported MQC classifier and behavior association. The relevant observable outcome here is: Matched voice flows receive the configured treatment and unrelated traffic remains within its approved class.

The restricted destination is denied and authorized comparison traffic succeeds in a controlled test. That observation would support a separate case of ACL packet filter rule order and direction, not verification of the reported MQC classifier and behavior association. The relevant observable outcome here is: Matched voice flows receive the configured treatment and unrelated traffic remains within its approved class.

Policy-based routing can select forwarding based on packet attributes before conventional destination routing. This explains Policy-based routing for approved application, but not why the issue in this scenario occurs. The specific mechanism is: Huawei MQC separates class identification, traffic behavior and applied traffic-policy binding.

Higher BGP LOCAL_PREF is normally preferred within an AS and is not advertised to external AS peers. This explains BGP local preference internal egress, but not why the issue in this scenario occurs. The specific mechanism is: Huawei MQC separates class identification, traffic behavior and applied traffic-policy binding.

AS_PATH length is one BGP best-path criterion after important preceding preferences. This explains BGP AS_PATH length egress choice, but not why the issue in this scenario occurs. The specific mechanism is: Huawei MQC separates class identification, traffic behavior and applied traffic-policy binding.

An Established session with received prefixes can still leave routes unusable if next hops are unresolved. This consequence is associated more directly with iBGP next-hop reachability. For the stated conditions the immediate exposure is: Defining a classifier without applying the intended policy gives a false impression of active QoS.

Ignoring community-scoped policy can expose unintended network reachability. This consequence is associated more directly with BGP community filtering in routing control. For the stated conditions the immediate exposure is: Defining a classifier without applying the intended policy gives a false impression of active QoS.

Untrusted DSCP markings can let ordinary traffic consume scarce high-priority capacity. This consequence is associated more directly with QoS DSCP remarking and trust. For the stated conditions the immediate exposure is: Defining a classifier without applying the intended policy gives a false impression of active QoS.

Compare received candidate paths, LOCAL_PREF values, route policy and installed best path. This check would diagnose BGP local preference internal egress, but would not first isolate the described MQC classifier and behavior association. Begin with: Inspect traffic classifier, behavior, policy association, interface binding and match counters.

Check route tags, source protocol, import-route statements and re-advertisement history. This check would diagnose Redistribution route tag loop avoidance, but would not first isolate the described MQC classifier and behavior association. Begin with: Inspect traffic classifier, behavior, policy association, interface binding and match counters.

Inspect rejected UPDATE attributes, local AS number and AS_PATH on each advertised prefix. This check would diagnose BGP AS_PATH loop detection, but would not first isolate the described MQC classifier and behavior association. Begin with: Inspect traffic classifier, behavior, policy association, interface binding and match counters.

Add supported PBR next-hop reachability tracking or failover logic consistent with service requirements. This is a legitimate response to the different issue of PBR failed-next-hop fallback, but it does not correct the specific MQC classifier and behavior association condition reported. The relevant technical distinction is: Huawei MQC separates class identification, traffic behavior and applied traffic-policy binding.

Reorder or narrow the route-policy nodes so specific restrictions are evaluated before general permits. This is a legitimate response to the different issue of Route-policy permit and deny node order, but it does not correct the specific MQC classifier and behavior association condition reported. The relevant technical distinction is: Huawei MQC separates class identification, traffic behavior and applied traffic-policy binding.

Design intended reflector cluster IDs and client relationships with loop-prevention attributes intact. This is a legitimate response to the different issue of BGP route reflector cluster loop prevention, but it does not correct the specific MQC classifier and behavior association condition reported. The relevant technical distinction is: Huawei MQC separates class identification, traffic behavior and applied traffic-policy binding.

A supported Level-1 adjacency forms only between appropriate same-area neighbors and routes appear correctly. That observation would support a separate case of IS-IS level-1 area adjacency, not verification of the reported MQC classifier and behavior association. The relevant observable outcome here is: Matched voice flows receive the configured treatment and unrelated traffic remains within its approved class.

The neighboring AS picks the intended ingress where the MED comparison is applicable. That observation would support a separate case of BGP MED same-neighbor-AS comparison, not verification of the reported MQC classifier and behavior association. The relevant observable outcome here is: Matched voice flows receive the configured treatment and unrelated traffic remains within its approved class.

A test prohibited prefix is rejected while authorized matching routes retain expected attributes. That observation would support a separate case of Route-policy permit and deny node order, not verification of the reported MQC classifier and behavior association. The relevant observable outcome here is: Matched voice flows receive the configured treatment and unrelated traffic remains within its approved class.

BGP MED is a nontransitive hint often compared among paths learned from the same neighboring AS. This explains BGP MED same-neighbor-AS comparison, but not why the issue in this scenario occurs. The specific mechanism is: Huawei MQC separates class identification, traffic behavior and applied traffic-policy binding.

QoS differentiation depends on packet classification, marking trust and congestion-aware queuing. This explains QoS DSCP remarking and trust, but not why the issue in this scenario occurs. The specific mechanism is: Huawei MQC separates class identification, traffic behavior and applied traffic-policy binding.

Ordered route-policy matching stops at an applicable node and its permit or deny behavior. This explains Route-policy permit and deny node order, but not why the issue in this scenario occurs. The specific mechanism is: Huawei MQC separates class identification, traffic behavior and applied traffic-policy binding.

An attribute-only change without forwarding checks can create a silent blackhole. This consequence is associated more directly with Route-policy next-hop and preference application. For the stated conditions the immediate exposure is: Defining a classifier without applying the intended policy gives a false impression of active QoS.

Changing DNS alone cannot repair missing DHCP relay or unreachable server paths. This consequence is associated more directly with DHCP relay to remote server. For the stated conditions the immediate exposure is: Defining a classifier without applying the intended policy gives a false impression of active QoS.

Assuming MED always overrides LOCAL_PREF or applies across every AS can misroute traffic. This consequence is associated more directly with BGP MED same-neighbor-AS comparison. For the stated conditions the immediate exposure is: Defining a classifier without applying the intended policy gives a false impression of active QoS.

Inspect route-policy node IDs, if-match criteria, apply clauses and hit counts. This check would diagnose Route-policy permit and deny node order, but would not first isolate the described MQC classifier and behavior association. Begin with: Inspect traffic classifier, behavior, policy association, interface binding and match counters.

Inspect route communities on received paths and the community-filter and route-policy match results. This check would diagnose BGP community filtering in routing control, but would not first isolate the described MQC classifier and behavior association. Begin with: Inspect traffic classifier, behavior, policy association, interface binding and match counters.

Inspect peer reflect-client settings, received and advertised routes and RR topology. This check would diagnose BGP route-reflector client scoping, but would not first isolate the described MQC classifier and behavior association. Begin with: Inspect traffic classifier, behavior, policy association, interface binding and match counters.