I’ve got a Conduit AEP running with a prepaid 3G sim card. I’m find the gateway drops off TTN after a day or two (still testing exact duration) and I need to cycle power on the Conduit to get it running again.
Has anybody experienced this or have any troubleshooting tips? I’ve already cycled the power today but the next time it drops off I’ll log on to it to check the services.
The cellular connection may be reset by the carrier resulting in the ppp0 interface IP address being changed.
Depending on the packet forwarder used, the autoquit_threshold for Kersing or Semtech packet fowarder can be used to detect the loss of connectivity and quit the process.
If an angel or monitor process is used, the packet forwarder should be restarted automatically and reconnect.
@jreiss does the autoquit work for Kersing’s MP packet forwarder?
without having to change anything else apart form adding an autoquit line in the global_conf.json file?
Does the AEP watchdog restart kersing’s packet forwarder?
(sorry if these are stupid questions, I have no idea how these things work)
regards
Paul
edit to add: or would you have a simple monitor process example? I have three of these in field with this problem (as in this thead Multitech - TTN connection lost after bad connectivity) and they are dropping connections pretty frequently, and farmers are relying on these for operations on theirs farms.
If you used a recent installer the packet forwarder will automatically restart if it exits, installed scripting will take care of that.
Auto quit works for Semtech transport only. TTN protocol reconnects itself. There is an issue in the current builds where TTN protocol reconnects fail that I’m working on resolving, I’ve tracked the issue to one of the libraries used, a possible fix is being tested.
thanks guys, heading out to one of my gateways (300km away) on Friday. (can’t be accessed remotely at the moment)
If you have anything you have confidence in by then, happy to give it try
cheers
Paul
We have 2 gateways using LTE “prepaid” sims that have the same problem. We are using the implementation of @kersing of mp_pkt_fwd.
I still doubt if this parameter is server specific or a general parameter for all Semtech UDP servers configured in the gateway.
The parameter is global, the counter is for each gateway. So if you set it to three and one of the gateways misses one and the other two PULL_ACK packets nothing happens. If one server misses three PULL_ACK packets the packet forwarder will exit.