Freeze one chassis, port, call and symptom
Record model, V2 label, firmware, MAC address, power and network path, physical FXS port, analog endpoint, cable or RJ21 pair, profile, SIP User ID, Authenticate ID, registration row, hunt group, PBX extension and one failed call direction. Preserve timestamps and redact secrets.
Classify the first divergence: wrong physical pair; endpoint off-hook or electrically incompatible; port assigned to the wrong profile; authentication or registration failure; registered identity differs from the dialled destination; hunt policy selects another port; outbound identity is rejected; signaling succeeds but DTMF, fax or media fails; failover or provisioning changed effective state.
Prove the analog port and pair first
Use the gateway status view to compare the numbered port and Hook state while lifting one known handset. If the wrong row changes, repair the jack, patch panel or RJ21 pair map before touching SIP. Confirm local dial tone, ringing and disconnect behavior on that port.
Do not bridge an outside analog line into an FXS port: this family supplies subscriber-side analog service and has no FXO port. Isolate cable length, splitters, REN load and attached fax or alarm equipment when electrical behavior is inconsistent.
Compare effective profile, identity and registration
Read the effective per-port profile, SIP User ID, Authenticate ID, masked credential state, SIP server and registration status. Then compare the PBX or provider account that should own the call. A Registered status can belong to the wrong account or profile.
If outbound calls are allowed without registration, do not use that success to prove inbound reachability. If Routing ID is configured, compare From and Contact presentation separately from the credentials used to register and authenticate.
Trace hunting and PBX routing independently
For inbound calls, compare the Request-URI or assigned number, upstream route, target SIP account and gateway hunt group. Check member order, linear or circular mode, active calls and Active Port Only versus All Ports registration. A hunt group selects an eligible port; it is not a simultaneous ring group.
For outbound calls, compare digits sent by the analog endpoint, dial plan, profile, presented identity and PBX route. Change one layer at a time and repeat the same call.
Correlate server switching, provisioning and firmware
If the fault is intermittent, retain profile status, keep-alive result, primary or failover selection, Prefer Primary setting, DNS answers and the transition time. A device choosing another server does not by itself prove upstream failure.
Compare intended provisioning with fetched, applied and effective configuration. Verify the V2 firmware filename and release-note path against the chassis label. Respect documented downgrade restrictions and do not cross-load V1 and V2 firmware as a recovery shortcut.
Verify one corrected call before widening the test
After the narrow correction, prove the same physical port changes hook state, holds the intended profile and identity, reaches the expected registration state and completes one authorized call in the failed direction. Then test DTMF, caller ID, fax mode, concurrent hunting and failover separately.
Record before-and-after port, profile, registration and routing evidence. No gateway, analog endpoint, cabling, RJ21 assembly, PBX, profile, credential, registration, hunt group, provisioning job, firmware image, fax, DTMF, call, media stream, power event or failover was inspected or tested.
Sources & applicability
Primary references for the technical details above. Operational examples and planning checklists are VoIP.info editorial guidance.
Examples require adaptation to your topology. No live PBX or hardware testing is claimed.