OTAA sessions and Conduit restart

[Guy]

Hi

I am trying to understand what happens if a Conduit running a local instance of the LoRaWAN network server restarts and there are a number of devices that have joined using OTAA.

My initial testing shows that the Conduit ignores the devices unless they re-join.

My guess is that if the devices were set to ABP then the sessions would survive a restart, but that requires a whole lot of other management.

Can anyone tell me what exactly is causing this and if there is any way around the issue ?

(My devices will eventually rejoin after a couple of failed LinkChecks but it may take a day or more for them to do that.)

Thanks

[Michael Mitchell]

Not sure if this is your issue, but if you are not using a static whitelist, and you are doing device add operations, you need to run database backup after adding the devices, or they may not be in the database when it restarts.

If you do this, you should not have to rejoin when the Conduit restarts.

• This reply was modified 5 years ago by Michael Mitchell.

[Guy]

Hi Michael
thanks, but I have added the devices on the WebUI so they are in the database

[Jason Reiss]

When you say “restart” the Conduit are you doing a graceful shutdown and restart or pulling the power cord?

When the power is cut recent sessions may be lost as they have not been written to flash.

The database is held in RAM and backed up to flash periodically. The default period is hourly and can be reduced to ten minutes. This process reduces the flash wear. The trade-off is loss of session if power is cut. On normal restart or reboot the database is automatically backed to flash.

Backup command info
http://www.multitech.net/developer/software/lora/lora-network-server/

Backup interval info
http://www.multitech.net/developer/software/aep/lorawan-aep/lorawan-network-settings/

[Guy]

Hi Jason

Thanks, that was the issue. I was pulling the plug to simulate a power failure but had not had the device running long enough since the join for the DB to backup.

I assume that whilst the device was in the whitelist the latest network keys had not yet been saved so the messages were rejected.

[Rob]

Guy
I am having a similar problem (and tearing my hair out in the process)

I am running the Conduit AEP in Network Server mode.

I have manually added all 16 No. LoRa End devices to the “END DEVICES” list on the dashboard (Devices tab). I then Save & Restart to save the details to the RAM.

All LoRa end devices (Elsys ELT-2-HP & Sensing-Labs) have the gateway keys configured to them so all this is 100% okay)

I spend 2hrs travelling around my site activating the end devices which send join requests and all is okay. I see all 16 devices in the “Sessions” list.

When I look a few hours later (without any power outage) I have lost them all. For “last seen” it reads “unknown”.

I have tried this several times for a week but still the same.

Should I try ABP instead of OTAA?

PS…ARGHH

• This reply was modified 4 years, 11 months ago by Rob.

[Steve Kovarik]

In Conduit AEP configured as LoRa Network Server, to add OTA endpoints
select “LoRaWAN” tab, then “Key Management” tab and “Add New” for
local end device credentials.

To add ABP endpoints select “Devices” tab and “Add New” for End devices and Sessions.

[Guy]

Ok, that is confusing.

I have just been using OTA and did not realise the difference in how you configure OTA versus ABP devices. So I guess I must have done it correctly.

You say that they joined ok though ?

You could try getting one device working reliably on your desk first and then run around and get the others working.

[Author]

Posts