An AMQP error occurred (condition='amqp:internal-error') during join requests about iotedge-lorawan-starterkit

[gerfen]

@fbeltrao Thanks. I'll close this issue after I've confirmed that I can connect more than 100 devices sometime later this afternoon my time.

[ronniesa]

This is expected because by default IoT Edge has a maximum of 100 devices concurrently connected. You can increase this settings on the edgeHub module with the env variable: "MaxConnectedClients".

Depending on the type of the lora device you can optimize for latency and downstream msgs (keeping the connection open for a long time, our default is 1 day) or for sensors density where after every lora message we close the connection to iot hub (this is good for something like parking sensors that are mostly, if not only upstream, and are not high frequency in sending msgs). You control this per device in specifying "KeepAliveTimeout" desired twin: https://github.com/Azure/iotedge-lorawan-starterkit/blob/master/Docs/quickstart.md#optional-device-properties

You could expect on an Intel gateway to have 500 or even up to 990 active connections at the same time. On a Pi we tested with 250 concurrent.

[gerfen]

@ronniesa Thanks for the info. I'm running the AAEON AIOT-IP68 gateway so I think I should easily be able to support 500 devices. Just want to confirm that I set the "MaxConnectedClients" variable in deviceConfiguration.json?

"env": {
    "OptimizeForPerformance": {
          "value": "false"
    },
    "MaxConnectedClients": {
	  "value": "500"
    },
    "mqttSettings__enabled": {
           "value": "false"
    },
    "httpSettings__enabled": {
            "value": "false"
    },
    "TwinManagerVersion": {
            "value": "v2"
    }
},

I've deploy 110 temperature and humidity sensors which are reporting every 5 minutes. If I understand correctly I think I should stay with the default KeepAliveTimeout for the device twins for these devices?

[ronniesa]

Correct.

If the devices are upstream only you could further optimize by setting the device twin “Downlink” to false.

[gerfen]

Would setting "Downlink" to false prevent sending C2D messages for setting things like reporting interval?

…

--Michael

________________________________ From: ronniesa <[email protected]> Sent: Friday, April 12, 2019 10:18:10 AM To: Azure/iotedge-lorawan-starterkit Cc: Michael Gerfen; Author Subject: Re: [Azure/iotedge-lorawan-starterkit] An AMQP error occurred (condition='amqp:internal-error') during join requests (#184) Correct. If the devices are upstream only you could further optimize by setting the device twin “Downlink” to false. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub<#184 (comment)>, or mute the thread<https://github.com/notifications/unsubscribe-auth/AFCS5Mb3e0T_xOp4pAQ9gN3_0LkzUICnks5vgL_SgaJpZM4cqxnz>.

[fbeltrao]

Yes, setting Downlink=false will prevent the device from receiving cloud to device messages.

[gerfen]

@ronniesa and @fbeltrao I've confirmed that I can register more than 100 devices with my gateway. Thanks for the assistance!

Closing the issue.