Skip to content

Latest commit

 

History

History
198 lines (139 loc) · 8.01 KB

Troubleshooting.md

File metadata and controls

198 lines (139 loc) · 8.01 KB
Raw

Troubleshooting Instructions

Overview of recent errors and warnings is provided in the server Event List. Details can be found in the server logs.

Depending on your installation the server logs are stored in /var/log/lorawan-server, /var/lib/lorawan-server/log or just lorawan-server/log. By default three log files are provided: debug.log, error.log and crash.log. The log messages contain date and time, severity (debug, info, notice, warning, error), process ID and a description.

For example:

2017-03-05 02:14:57.148 [info] <0.361.0> Gateway <<1,39,235,255,255,176,23,196>> at {192,168,0,1}:47002

Binary identifiers are often shown in the Erlang binary notation, enclosed in << >>. This is simply a list of bytes in decimal notation. For example, <<1,39,235,255,255,176,23,196>> means 01-27-EB-FF-FF-B0-17-C4.

By default no messages are shown on the console. To see the debug messages you need to open the server configuration in lorawan-server/releases/<VERSION>/sys.config, uncomment (remove the % character) from the {lager_console_backend, debug} line and restart the server.

Common Errors

No message in log file

The log shall always contain at least an [info] record Gateway <<MAC>> at {IP}:PORT. If this is not listed, the gateway is down or the network connectivity is broken.

This may be due to:

  • Gateway misconfiguration. Check the gateway configuration includes a correct server IP and port (by default 1680).
  • Firewall misconfiguration. Check the server firewall does not block the port 1680. See Installation Instructions for configuration guidelines.

No downlink frames delivered

This may be because:

  • Gateway configuration error. Verify you set correctly the TX Chain in your Infrastructure configuration.
  • Gateway error. Check for any error in your gateway log file.
  • The device and gateway are using inconsistent private/public network mode setting.
  • The device did not listen on the channel (frequency) used by the server. Verify your device correctly listens in the right RX1/RX2 window. The RX2 frequencies and data rates are provided in the regions config parameter.

Some devices like the Arduino LMIC may have problems receiving downlinks in the RX1 window. To bypass this you can set the TX Window of your Device or Node to `RX2'.

Join request received, but no further uplink frames

This is usually because the downlink frame including the Join response was not sent or received by the device. See No downlink frames sent above.

unknown_mac

The gateway MAC is not correctly configured. Open the Web Administration, go to the Infrastructure-Gateways list and check the gateway MAC is correctly listed.

unknown_deveui

The over-the-air activated (OTAA) device DevEUI is not correctly configured. Open the Web Administration, go to the Devices list and check the device EUI is correctly listed.

unknown_devaddr

The device DevAddr activated-by-personalization (ABP) is not correctly configured. Open the Web Administration, go to the Nodes list and check the device address is correctly listed.

If you see strange DevAddr numbers from devices that you don't know, there may be a second network near you. You can add devices from this network into the list of Ignored Nodes.

bad_mic

The Message Integrity Check (MIC) of a received frame has failed.

  • If this appeared only once this was due to a transmission error.
  • If this happens periodically with the same ID, the device is not correctly configured.
    • When this is related to a 8-bytes long DevEUI, open the Web Administration, go to the Devices list and check the AppKey.
    • When this is related to a 4-bytes long DevAddr, go to the Nodes list and check the NwkSKey.

fcnt_gap_too_large

The device sent an unexpected frame counter FCnt. This may be because:

  • The device is activated-by-personalization (ABP) and it did reset recently.
  • The device was unreachable for a very long time.

If this is an exceptional case, go to the Nodes list and manually update the FCnt Up to the FCnt number.

To allow ABP devices to freely reset set the FCnt Check to Reset on zero, but please note this weakens LoRaWAN security a bit. It is recommended to use over-the-air-activation (OTAA) instead.

repeated_reset

No frames were received since last OTAA join or last ABP reset. This is just a warning. It may be because:

  • No downlink frames delivered (see above)
  • Faulty device is periodically re-starting

not_semtech_mote

The server is configured to use the semtech-mote application, but your device sent some other data. Make sure your firmware uses the supported format.

downlink_missed

Confirmed downlink was sent, but the device indicated it did not received it. Depending on the application logic the downlink may be retransmitted.

downlink_expired

Confirmed downlink was superseded before it could be transmitted to the device. See the D/L Expires setting of the corresponding Handler.

prerequisite_failed

This is reported when the lorawan-server is started with older Erlang/OTP. At least 19.0 (or later) is required.

connector_disabled

The server is configured to use a connector, which is disabled. Set the Enabled flag of your Backend connector configuration.

Unknown downlink target

This is reported when a Backend connector doesn't know to what device a downlink shall be sent. You have two options:

  • Define a "devaddr" or "deveui" field in the JSON structure, or
  • Define the "Received Topic" as a template "in/{devaddr}", which causes the server to parse the topic and extract the DevAddr from there.

Unknown element in JSON received from the Lorank8 gateway

The problem is the Lorank8 proprietary message format. In your gateway config you likely have stat_format set to idee_concise or idee_verbose. You need to change stat_format to semtech to get this working.

Browser reports NS_ERROR_NET_INADEQUATE_SECURITY

You are using old version of Erlang that does not support the newest ciphers required by HTTP/2. Erlang 20 is recommended for TLS/SSL.

As a quick fix, in Mozilla you can disable the checks in about:config, by setting network.http.spdy.enforce-tls-profile to false.

Lost admin password

If you forgot your admin password

  • connect to the cluster via remote shell by erl -sname test -remsh lorawan@<hostname>
  • delete the user database by mnesia:delete_table(user).
  • restart the server to recreate the user database with the default admin password

Alarms

Server alarms are indicated in the web-admin Dashboard.

system_memory_high_watermark

This indicates that the system has less than 20% of free memory.

disk_almost_full

This indicates there is less than 20% of free disk space.

Issue Reporting

If this guideline didn't help you to solve your problem, feel free to report a new issue. For the better understanding of the issue you are encouraged to provide:

  • Distribution type: binary or built from source.
  • Hardware type: PC Server, Just PC, MacBook, MacServer, Raspberry Pi1/2/3/x, Odroid, etc.
  • OS type and version. Also, add the output of the uname -a command if you are running on a sort of a Unix/Linux/MacOS system.
  • Your Erlang/OTP version. Can be obtained by running erl at the command line: please provide all the text before 1> line.
  • If you are experiencing problems with the web-interface or seeing lots of http_error lines in the log , provide npm -v and node -v output too.

Get ready to provide your server's debug.log file and, in hard cases, the Wireshark network capture file.