Get device-session error with multiple network server

We’re trying to run multiple network servers with a single app server and gateway bridge, and see these error messages from the NS that aren’t provisioned for the device.

gatewaybridge_1            | time="2018-01-18T21:52:47Z" level=info msg="gateway: received udp packet from gateway" addr= protocol_version=1 type=PushData 
gatewaybridge_1            | time="2018-01-18T21:52:47Z" level=info msg="gateway: rxpk packet received" addr= data="gM2OuIwASgABCtojKaWprqbWnSCvPqBXL2TKAQC/" mac=00800000a0000153 
gatewaybridge_1            | time="2018-01-18T21:52:47Z" level=info msg="gateway: sending udp packet to gateway" addr= protocol_version=1 type=PushACK 
gatewaybridge_1            | time="2018-01-18T21:52:47Z" level=info msg="backend: publishing packet" topic="gateway/00800000a0000153/rx" 
lora-ns-us902_928_1        | time="2018-01-18T21:52:47Z" level=info msg="backend/gateway: rx packet received" 
lora-ns-eu863_870_1        | time="2018-01-18T21:52:47Z" level=info msg="backend/gateway: rx packet received" 
lora-ns-eu863_870_1        | time="2018-01-18T21:52:47Z" level=error msg="processing rx packet error: get device-session error: device-session does not exist or invalid fcnt or mic" data_base64="gM2OuIwASgABCtojKaWprqbWnSCvPqBXL2TKAQC/" 

I’m assuming this is because both NS are listening to the same MQTT topic, is there a way to filter this so each NS is only listening for traffic it cares about? This is using AS 0.16.1 and NS 0.23.2.

The issue is indeed that both network-servers see all the data, including the data of devices that are not activated. I’m thinking of dropping this log, or changing it to debug maybe, as this is also logged when one of your gateways receive data from nearby devices (which can result in a lot of “errors”). Let me know what you think of this.

The more I think about it the more I think maybe this is a non-issue. Our particular problem is because we’re trying to support multiple bands at our main location for development and manufacturing, but in production deployment we should only need one NS per region so we wouldn’t really be seeing this.

I wouldn’t get rid of the error completely, I think it could be useful during debugging as an indicator that a NS or gateway is being overloaded with traffic from foreign devices.