Devices stopped to receive uplinks after v3 migration

Yesterday we decided to migrate to v3 version, chosen one of our applications and used windows CLI tool to generate and export devices JSON. All devices were activated by ABP.

After re-creating the devices in v3, we not only were not able to make them work but it also stopped working in v2 (other, untouched applications are still working). No live data for the devices both in v2 and v3, I have triple checked all IDs and keys. We have changed application name in v3 to more accurate, but I believe it’s not important?

Finally after checking through other threads tried to change ABP to OTAA for one device but still not lucky. Also, we didn’t migrate all gateways as I’ve read that they point to v2 ones anyway.

We’re not really experienced in TTN, in v2 it was very easy for us to setup the gateways/devices but we struggle here. Any help would be appreciated.

ABP traffc for V2 address ranges is not forwarded from V2 gateways to V3. Only V3 connected gateways will provide ABP traffic for V2 addresses in V3.

That’s right, the only things of importance for LoRaWAN traffic are DevAddr, NetworkSKey and AppSKey (for ABP) or DevEUI, AppEUI/JoinEUI and AppKey for OTAA where you will need to force a join.

It would help if you could elaborate a bit. What did you do? Is there any traffic in the V3 console for the OTAA device? Just stating it doesn’t work makes it hard for people to help.

About OTAA - I don’t have physical access to the devices right now, so I only changed the Activation Method in v2 and it generated the App Key. There is no traffic at all for the devices. Tried to send downlinks yesterday through v2, reconfiguring the devices, but didn’t help.

I know, but as I mentioned above, we are not very experienced with The Things Network. We don’t even know what informations would you need to help me in this case. And most important, we don’t know why v2 stopped working in the first place.

Did you use these instructions:

If you can’t get physical access then you will HAVE to get a v3 gateway setup and use the same DevAdr, NwksKey and AppsKey in the setup on the v3 console.

Just please make it clear for me:
How to Migrate Gateways from V2 to V3 - here is the advice to not migrate gateways to v3
How to Migrate Gateways from V2 to V3 - #10 by Jeff-UK - here is the advice to not delete gateways because of further problems with re-creating them

But I also used your advice to create v3 gateways (same as in v2) - status disconnected on all of them and no data received.- according above warnings, should I delete v2 gateways, or what is the correct step there?

Creating a gateway in V3 does not move the gateway to V3. For that to happen the configuration on the gateway needs to change and that is not controllable in the TTN console.
What model gateway are you using?

(And no, do not delete gateways in V2)

We are using Things Network indoor gateways.

They are not enabled for V3 YET.

The reason I said you’d have to setup a V3 gateway is because you said you’d moved all your devices to V3.

If you haven’t deleted/disabled them in V2, I’d disable them in V3 and quit (temporarily) whilst you are ahead.

If you have no physical access to the devices, how did you manage to change it to OTAA?

So you didn’t bother to read the rest of the thread where I posted that you have called out above??? :man_shrugging:

I understand we’ve made a falstart so I will go back to v2 till v3 will be ready. I changed to OTAA only in the console device settings, to look if it will make any change or display some errors as I preffer error than nothing at all.

They are not deleted/disabled in v2, but even when I deleted the devices from v3 they are still not calling in v2. Should I do something else or just wait for them to re-join v2 automatically?

ABP do not join so they won’t do that automatically.

Do you have the green light on your V2 TTIG?

What traffic does it show? Not status, status is broken in the V2 console, only traffic counts.

image
Device status in v2 is green but no data from nodes in range of devices, except for the Device friendly name CarbonCounter Sedgefield

As Things Network indoor gateways in V3 will not work and are marked as disconnected should they be deleted. Not that I can see an option for deleting them.
I’m reactivating the end devices (sensing-labs kit) using a magnet. The process appeared successful in that a solid green light for 2 seconds at appropriate time. Now wating for the V2 console to update status to green connected and traffic from end device.

An actual physical green light on the unit - that tells you the TTIG is connected and doesn’t have issues with Wifi or similar.

The online status is broken. I think I said that. Already.

Is this a light on the device? Or the console? And if it’s the console, which part of the console?

Which console? If the gateway console, then you should no rely on the green dot status as it’s not reliable / broken. This may have been mentioned.

Traffic on the console for both gateway and device is the only thing that counts. This may have been mentioned.

One end device has reconnected in v2 console, with occasional traffic. The other end-devices appear to be dumb. The gateways to which these end devices are linked are all visible and seen every hour, this suggests they are picking up traffic but I’m not seeing any traffic in console. Have checked settings against the commissioning file from supplier

If they were picking up traffic they would, rather obviously, show traffic in the console. Even traffic for “foreign” devices - ie not TTN or TTN but not registered on one of your applications.

How many devices do you have deployed?

Do you have a spare device to hand that you can take out to your TTIGs to check if they are actually online by virtue of traffic from device to application and out the other end, the only real test worth doing.

Have you changed the AppsKey in the v3 console so there is no chance they can be attempted to be processed by the v3 stack?

Have you reinstated the exact settings in the v2 console for the devices so they can be processed on that side?

I’ve deleted the Application API key in v3. Only one node is in the v3 app. I have two live nodes in my home on electric meters and two TTIGs. And several other previously activated nodes that were “flat-lining” but with traffic.
image
The app key is same as commissioning csv - node has been working for months

Sorry, rather like the Stack Overflow issue, my IQ wraps back round to being me being rather dumb.

Can you answer the questions - I can touch type but I still think about what I ask.

PS, the Application API key allow you to access the servers from outside of it, no relevance to the situation at hand. Maybe you mean Application EUI - you can delete devices safely and re-instate and you can delete applications but can’t reuse the id but can reuse the EUI (possibly, not tried it)

PSS, did I mention the Aspbergers / OCD - which means I’m a bit picky about the nomclementure.

Found a solution. Have to make a hard reset of the devices - tweezers across two points on the circuit board. Delete device from v2 application that is not working. For absolute certainty am making the hard reset again after deleting,
Next step add the devices to a new application in v2 (am using the same settings except for uniqueID)
After device activation, which works successfully am seeing device traffic