Problem with IV index udpate

RE: Problem with IV index udpate

Fabbros — Fri, 24 May 2024 13:51:08 GMT

I'm writing this down to let you know that, thank to the guys on github, I was able to fix the problem, see https://github.com/NordicSemiconductor/Android-nRF-Mesh-Library/issues/537

RE: Problem with IV index udpate

Hung Bui — Fri, 21 Oct 2022 11:07:54 GMT

Hi Fabbros,
I would suggest to continue the discussion with Philips on github.
What we are seeing in your log is that the SNB beacon was rejected because it was going straight from Index 1 active (In progress) state to Index 2 active (In progress) and that's not allowed.
It must go from Index 1 active to Index 2 normal before go to Index 2 active.

What need to be checked is that you have test mode set up on all the nodes. And you go step by step, first update form Index 0 to index 1, make Index 1 normal. Then change index 1 to index 2. Make index 2 normal before sending the next one.

RE: Problem with IV index udpate

Fabbros — Thu, 20 Oct 2022 15:08:10 GMT

Here are the log and a screen record of the following operations:

connect to mesh (with IV 2)
try to read a generic_onoff_server status -> FAIL
try to set a generic_onoff_server status -> FAIL

devzone.nordicsemi.com/.../Record_5F00_2022_2D00_10_2D00_20_2D00_17_2D00_00_2D00_20.mp4

devzone.nordicsemi.com/.../get_5F00_set_5F00_generic_5F00_onoff_5F00_server.txt

devzone.nordicsemi.com/.../Record_5F00_2022_2D00_10_2D00_20_2D00_17_2D00_00_2D00_20.mp4

RE: Problem with IV index udpate

Hung Bui — Thu, 20 Oct 2022 14:22:05 GMT

Thanks for the info. We will try to reproduce the issue here.
In the mean time could you try to get any log from the app by connecting it to a computer ?

RE: Problem with IV index udpate

Fabbros — Thu, 20 Oct 2022 13:36:15 GMT

I work for a Home Automation Company that developed an app (both Android and iOS) which configures our mesh devices (based on Nordic). I'm talking about thousands client that are using our app. We started experiencing several cases where IV is stepped to 1, with the time passing by we are expecting to get cases with IV 2 very soon and that will be dramatic.

First time we saw the issue we were using and old android SDK which wasn't able to communicate with device with IV 1, so we updated the SDK to fix that. Now we're doing some digging into the IV procedure and find out that if the IV step to 2 the app stop working again. To do the test we modify our devices enabling the IV test flag (to skip the canonical 96 hours) and force one of them to trigger the IV procedure from 0 to 1 and then from 1 to 2.

To verify the current IV status, and the current status of IV procedure (Normal vs update) we use nRF Connect because it shows the devices SNB Advertising.

RE: Problem with IV index udpate

Hung Bui — Thu, 20 Oct 2022 13:17:43 GMT

Hi Fabbros,

Thanks for the info. The issue is seem to be on the app side.
If you disconnect and connect to already provisioned proxy would you be able to communicate with them ?
We will try to reproduce the issue here and let you know what we find.

If you can list exact step on how you trigger the IV index update, how you verify that it's on Normal state, etc it would be great.

RE: Problem with IV index udpate

Fabbros — Thu, 20 Oct 2022 12:29:48 GMT

Hi Hung Bui,

thank you for taking over.

I tried setting the Generic_OnOff_Client of device A to publish to the Generic_OnOff_Server of device B and they work fine also with IV 2.

The app stop working after IV Update state change to Normal operation. After that, the only thing I can do is connecting to the proxy. Everything else seem not to work (I checked: apply binding, read/set state to a Generic_OnOff_Server)

RE: Problem with IV index udpate

Hung Bui — Wed, 19 Oct 2022 11:02:50 GMT

Hi Fabbros,
I'm sorry for the late response. Håkon is on a business trip so I will take over the case.

I will check with the app team to see if there is any issue with the app.

But in the mean time to isolate the problem to see if it's the issue only on the app or it also an issue with the mesh firmware, could you let me know after you do IV update from 1 to 2 , other devices in the network can still communicate to each other ? For example if you have a client and a server (both running on nRF5 device) can they communicate to each other when doing IV update ?

At which IV Update state (Normal/In Progress) do you see the app on the phone stopped responding ? When the issue happen, can you control the node that already provisioned to the mesh network ?

RE: Problem with IV index udpate

Fabbros — Wed, 19 Oct 2022 08:00:19 GMT

Is there anyone who can help me?

RE: Problem with IV index udpate

Fabbros — Wed, 12 Oct 2022 12:41:38 GMT

Hi Elfving,

of course, I did all the steps described above with and without IV Test Mode flag enabled.

I ran all the steps again today just to be sure.

RE: Problem with IV index udpate

Elfving — Tue, 11 Oct 2022 12:46:15 GMT

Hello Fabros,

Did you have a look at the suggestion by Philip on Github?

Regards,

Elfving