Field Test Data Processor extension — Device Overview and CSV export empty despite working uplink pipeline (GET /ext/db returns 500)

LPWAN Industrial Gateways / Concentrators
1

Hardware/Firmware:

  • Field Tester: RAK10701-Plus, firmware v1.0.4
  • Gateway: RAK7268V2, WisGateOS_2.2.16_RAK
  • Extension: Field Test Data Processor, latest from Extension Gallery
  • LNS: AWS IoT Core for LoRaWAN

Issue

The extension is confirmed receiving and correctly processing uplinks — including Location Label data — but the Device Overview page and CSV export remain completely empty, and the Field Tester’s own screen shows no data either.

What’s confirmed working:

  • Gateway connects to AWS IoT Core via Basic Station (CUPS + LNS)
  • Extension connects via MQTT (8883, TLS), subscribes, and receives uplinks continuously (~every 10-12 sec)
  • Field Tester joins and transmits normally
  • Setting a Location Label (e.g., “TEST1”) is correctly decoded by the extension — confirmed via payload match — and logs Start Label Test... for every labeled uplink over many minutes

What’s not working:

  • Device Overview: no data ever appears
  • CSV export: empty
  • Field Tester’s on-device screen: no uplink/performance data shown

Key log evidence:

Wed Sep 16 02:19:53 2026 user.info com_rak_mapper[27390]: MQTT client connected
Wed Sep 16 02:19:53 2026 user.info com_rak_mapper[27390]: Topic subscribed successfully
Wed Sep 16 02:19:57 2026 user.debug restify: [500, -1] Internal Server Error
Wed Sep 16 02:19:57 2026 user.info restify: 127.0.0.1 - - [16/Sep/2026 02:19:57] "GET /ext/db HTTP/1.0" 500 -
Wed Sep 16 02:20:02 2026 user.info com_rak_mapper[27390]: Received uplink message: {'MessageId': '[redacted]', 'WirelessDeviceId': '[redacted]', 'PayloadData': 'VEVTVDEAgA==', 'WirelessMetadata': {'LoRaWAN': {'ADR': False, 'Bandwidth': 125, 'ClassB': False, 'CodeRate': '4/5', 'DataRate': '0', 'DevAddr': '[redacted]', 'DevEui': '[redacted]', 'FCnt': 19, 'FOptLen': 1, 'FPort': 5, 'Frequency': '905100000', 'Gateways': [{'GatewayEui': '[redacted]', 'Rssi': -29, 'Snr': 12.75}], 'MIC': '[redacted]', 'MType': 'ConfirmedDataUp', 'Major': 'LoRaWANR1', 'Modulation': 'LORA', 'PolarizationInversion': False, 'SpreadingFactor': 10, 'Timestamp': '2026-09-16T02:20:02Z'}}}

GET /ext/db — the endpoint that appears to feed the Device Overview/export UI — returns 500 right after an MQTT reconnect, and is reproducible. Uplink processing continues normally afterward, but the stored dataset behind that endpoint appears broken or uninitialized.

Question: Has anyone seen /ext/db return a 500 on this extension? Is there a known fix (extension reinstall, DB reset, specific firmware pairing)? Happy to provide more logs if useful.

2

Updates:

Follow-up to my earlier post (no responses yet) — found a more specific, reproducible pattern that may point to the actual root cause.

The extension’s MQTT client is disconnecting and reconnecting after every single uplink/downlink cycle, correlated specifically with sending the downlink response. This is consistent across two separate days and two separate process restarts (different PIDs), so it appears to be a repeatable behavior of the extension itself, not a one-off network blip.

Log pattern (repeats every ~10-12 sec, matching the Field Tester’s uplink interval):

[Day 1, PID 32722]
Wed Sep 16 00:42:32 2026 user.info com_rak_mapper[32722]: MQTT client connected
Wed Sep 16 00:42:32 2026 user.info com_rak_mapper[32722]: Topic subscribed successfully
Wed Sep 16 00:42:41 2026 user.info com_rak_mapper[32722]: Received uplink message: {...}
Wed Sep 16 00:42:41 2026 user.info com_rak_mapper[32722]: Sent downlink message to [DevEUI]: {'deviceID': '[redacted]', 'confirmed': False, 'fPort': 10, 'data': 'djEuMS4wAA=='}
Wed Sep 16 00:42:41 2026 user.info com_rak_mapper[32722]: MQTT client disconnected with reason code: 1

[Day 2, PID 27390 — same pattern, different process instance]
Wed Sep 16 23:02:08 2026 user.info com_rak_mapper[27390]: Received uplink message: {...}
Wed Sep 16 23:02:08 2026 user.info com_rak_mapper[27390]: Sent downlink message to [DevEUI]: {'deviceID': '[redacted]', 'confirmed': False, 'fPort': 10, 'data': 'djEuMS4wAA=='}
Wed Sep 16 23:02:08 2026 user.info com_rak_mapper[27390]: MQTT client disconnected with reason code: 1
Wed Sep 16 23:02:10 2026 user.info com_rak_mapper[27390]: MQTT client connected
Wed Sep 16 23:02:10 2026 user.info com_rak_mapper[27390]: Topic subscribed successfully

The disconnect happens immediately after “Sent downlink message” every time — never after an uplink that doesn’t trigger a downlink response. This strongly suggests the downlink-send code path is tearing down the MQTT session as a side effect, rather than reusing a persistent connection.

  • Uplinks and Location Labels are being received and correctly decoded by the extension on every cycle (confirmed via payload match to the exact label set on the device)
  • But Device Overview, CSV export, and the Field Tester’s own on-device display all remain empty
  • The previously reported GET /ext/db → 500 error occurs right after an MQTT reconnect — plausibly a consequence of this same connection churn destabilizing whatever process persists data to the extension’s local database

Question: Is this disconnect-after-downlink behavior expected/by design, or a bug? Is there a way to confirm whether the reconnect is completing cleanly enough for the LNS to actually deliver the downlink, or whether it’s dropped in the process? Any known fix or workaround (extension version, config option) would be appreciated — happy to provide further logs.