Hi all,
I'm validating an SRv6 EVPN E-Line (XConnect) service on NCS5501-SE (IOS XR 7.8.2) using SRv6 micro-segment (uSID) with Flex-Algo 128 locators.
Traffic arrives on the SRv6 side and appears to decapsulate correctly (CEF and NPU both show forwarding),
but no traffic egresses the AC interface (Te0/0/0/11.1000).
■ Setup Summary
Ingress (SRv6) : Hu0/0/1/2
Egress (AC) : Te0/0/0/11.1000 (dot1q 1000)
Local SID (uDX2) : fd00:1:238:e002::/64
Context { evi = 65000, eth-tag = 1 }, State = InUse
Locator : fd00:1:238::/48 (micro-segment / unode psp-usd, Flex-Algo 128)
XConnect : UP ? both AC and EVPN segments Bound
EVPN BGP : Type-1 and Type-4 routes received properly
■ Observed Symptom
1000 pps received on Hu0/0/1/2
CEF resolves fd00:1:238:e002::/64 -> Te0/0/0/11.1000
NPU / EPNI counters increase
However, output rate on Te0/0/0/11 remains 0
In show l2vpn forwarding detail, EVPN segment shows received = 0 / sent increasing
■ Verification Results
show segment-routing srv6 micro-segment local-sid prefix fd00:1:e002::/48 internal
-> uDX2 InUse context { evi=65000, eth-tag=1 }
show cef ipv6 fd00:1:238:e002::/64 hardware ingress/egress
-> AC Interface resolved (Te0/0/0/11.1000)
show l2vpn forwarding counter location 0/0/CPU0
-> Byte Switched increasing
show controllers npu stats traps-all instance all location 0/0/CPU0
-> no SRv6 drops
show controllers npu stats counters-all instance all location 0/0/CPU0
-> packets reaching EPNI
show running-config interface Te0/0/0/11.1000
-> l2transport encapsulation dot1q 1000
■ Request for Advice
Could someone help identify what might be missing or which additional checks would clarify where the packet stops?
Possible areas I suspect:
Bridge-domain / L2FIB mismatch
- Could an implicit Bridge-ID be mismatched since no explicit bridge-domain is configured?
SRv6 -> EVPN handoff path
- Is a /64 SID with { evi, eth-tag } context valid for uDX2 under micro-segment mode?
- Could an overlap with the /48 Flex-Algo locator cause incorrect SID resolution or next-hop binding?
EPNI counters increase but no physical egress
- Are there known cases where rewrite profile or bridge-domain binding issues drop frames just before egress?
■ Additional Notes
CEF and RIB resolution look fine
NPU forwarding up to EPNI confirmed
Decapsulation appears complete, but packet is never handed to the AC output path
If any specific command output would help, please let me know and I'll share it.
Any insight or hint would be highly appreciated.
Regards,
