Jason ReissTo force
$ lora-query -x database backup
Otherwise it is backed up on interval, 1 hour default, 10 minutes min setting.
Jason ReissSet conduit to public mode with join delay 5 seconds.
Jason ReissWhat is the origin of the lora package?
Jason ReissIt may have been locked by the seller. I would contact them for instructions.
Jason ReissIs this a lorawan gateway for a lorawan network provider?
Jason ReissPlease open a case at http://support.multitech.com for assistance.
Jason Reiss.
Jason Reiss1. Everything seen in the UI is available through the RestAPI.
POST|GET|DELETE /api/lora/groups
2. Do you mean uplink/downlink from the end-device actions? Anything changing in the end-device from the network server will be sent in a downlink. Anything changed at the end-device will be seen in an uplink.
Jason ReissThis error happens on a session save.
[ERROR] SPIFFS_remove failed -10002
First any existing session files are removed. Don’t know they are all gone until the delete fails.
Then the new session file is save.
There was an issue where multiple session files were created and an old session would be restored.
Jason ReissWhat setting is used on Conduit? Private MTS or Private LoRaWAN
The setPublicNetwork input has changed from bool to int.
For mDot::setPublicNetwork(false || 0) => Private MTS
For mDot::setPublicNetwork(true || 1) => Public LoRaWAN
For mDot::setPublicNetwork(2) => Private LoRaWAN
Jason ReissDid you try:
$ opkg remove lora-network-server
$ opkg install lora-network-server_*.ipk
Jason Reiss(Net ID = crc64(network_name))
(Net key = cmac(passphrase, defaultKey))
The default session key is used with the passphrase as input to the CMAC.
2b.7e.15.16.28.ae.d2.a6.ab.f7.15.88.09.cf.4f.3c
Jason ReissDo have the correct antenna installed?
Jason ReissSo IMST shows -51 vs -104 dBm.
Gateway misses 20% according to the PER.
Looks to be a hardware issue.
Jason ReissThis works on mDot with latest mbed-os library after replacing mbed library with https://github.com/ARMmbed/mbed-os/
https://os.mbed.com/teams/Semtech/code/SX1272PingPong/
Edit: example does not support xDot
Jason Reiss.
Jason ReissIsmo,
Can you create a ticket at support.multitech.com?
Please upload full config from the Conduit and record network server logging at TRACE level there.
Jason ReissWe have not pushed an updated recipe yet. The tarball is available here if you want to update your recipe in mLinux.
http://www.multitech.net/downloads/lora-network-server_arm926ejste_2.1.5.tar.gz
Jason ReissDo you have similar results if the Conduit is set in packet forwarder mode and pointed to TTN?
Do you see the same channels being missed or is it random? For instance 868.3 MHz was missed above, is this channel always missed?
What is the SNR/RSSI of packets received at the gateway?
Can you provide output from conduit of the installed card?
$ mts-io-sysfs show lora/*
What is network server log output?
What is packer forwarder output?
It was once working with Conduit and now nothing?
It is not be expected for the device to be joined in both network servers via OTA.
Jason ReissJason ReisslibmDot libraries do not support this. The library uses lorawan to support authentication and encryption of the P2P connection.
An Sx1272 library is available.
https://os.mbed.com/teams/Semtech/code/SX1272Lib/#b988b60083a1
Jason ReissIs it OTA or ABP provisioned device?
If the OTA end device is joined to the ttn network server the conduit will not have session keys to authenticate and decrypt the packet.
Tcpdump of the uplink port 1780 should show packets received. Also network server log should show parsed and rejected packets.
Jason ReissA bug exists where the memory queue will not be cleared from the lora-query command. Likewise the database queue is not cleared when an MQTT message to clear the queue is issued. The next version, release pending, will have this issue fixed.
MQTT message to lora/EUI/clear and lora-query -x packet queue delete will empty both on the current release.
Jason ReissYes the application can stop the Join Accept from being sent based on EUI of the end-device.
A record will be entered into the database for the end-device that you may also want the application to clear out.
$ lora-query -x device delete EUI
As seen in the example below, the clear command blocks the packet_sent message. The second join_request is with the block disabled.
admin@mtcdt:/var/run/lora/1# mosquitto_sub -v -t lora/+/+
lora/00-00-00-00-00-00-27-10/join_request {"tmst":1150039354,"chan":8, ...
lora/00-00-00-00-00-00-27-10/packet_recv {"tmst":1150039354,"chan":8, ...
lora/00-00-00-00-00-00-27-10/joined {"remote_js":false}
lora/00-00-00-00-00-00-27-10/clear {"remote_js":false}
lora/00-00-00-00-00-00-27-10/join_request {"tmst":1167037612,"chan":7, ...
lora/00-00-00-00-00-00-27-10/packet_recv {"tmst":1167037612,"chan":7,...
lora/00-00-00-00-00-00-27-10/joined {"remote_js":false}
lora/00-00-00-00-00-00-27-10/packet_sent {"appeui":"16-ea-76-f6-ab-66-3d-80", ...
With Steve’s advice the end-devices can be left at AT+FSB=0 64 channel mode. They will try each FSB until they join a conduit. It will join in less than 16 attempts.
The whitelist is configured in the db.json file within a config.tar.gz uploaded to DeviceHQ. This list can be modified on a PC or extracted from a development Conduit and pushed down to production Conduits through DeviceHQ.
Otherwise an application can enforce a whitelist by sending MQTT message to lora/EUI/clear topic when the lora/EUI/joined message is seen for end-devices you do not want to join.
Jason ReissCould be caused by browser caching hmtl page. Try a hard reload of the page.
Jason ReissTry to save in network server mode, then switch to packet forwarder.
If that does not work try reset to default and reconfigure.
Jason ReissAt what datarate is the first uplink following the Join Request/Accept message being sent?
The ADR commands to set the Channel Mask, sent in the first downlink, should be instructing the device to keep the current datarate, it should only send a command to set to DR0 if the uplink used DR0.
Jason ReissSee Chapter 5 – Power Management
https://www.multitech.com/documents/publications/manuals/s000643.pdf
AT+WM controls wake-up condition
AT+WP controls wake-up pin
at+wp=?
AT+WP: (1-6)
OK
help at+wp
AT+WP: Wakeup GPIO pin of sleep mode (1-6) (default: WAKE, 1:UART1_RX), deep-sleep uses WAKE
OK
help at+wm
AT+WM: Wakeup mode, INTERRUPT uses WAKE as wake-up pin (0:INTERVAL,1:INTERRUPT,2:INTERVAL_OR_INTERRUPT)
OK
Jason ReissEdit: The current 2.0.19 network server implements LoRaWAN spec 1.0.2 with DR0-DR6.
If ADR is disabled the stack will accept the channel mask regardless of datarate.
The upcoming release will use 1.0.2rB.
Jason Reiss.
Jason ReissAn application can send the txparamsetupreq command using port 0.