[pre-sve] NotifyUpdateApplied fails if triggered before the node joins the fabric

[selissia]

Problem

When a node reboots after a successful OTA update the sending NotifyUpdateApplied might get triggered before the Node re-joins the fabric. In this case the command fails to be transmitted. This fails the TC-SU-2.6 test case

For example, in case of Thread transport re-joining the Thread network takes some time, see attached log snippet.

Proposed Solution

OTARequestor class should delay sending NotifyUpdateApplied until the device has joined the fabric.

[selissia]

tc-su-26-dut-log.txt

[carol-apple]

I think it'd be tricky to figure out how long to delay. There are no timestamps in your log snippet. How long between the first failed Advertise operational node 2BF32B802FF0AE27-0000000000000002 vs the second one which was successful?

[selissia]

Several seconds, I don't have a more specific info than that. But this will vary between devices and network conditions anyway.
I meant "delay" in a more general sense, as in "don't attempt NotifyUpdateApplied until the device has joined the fabric" -- how exactly to accomplish that I'm not yet sure.

[marius-alex-tache]

Hi! I'm seeing this after trying to increase OT baud rate to 921600. Seems like the call to HandleDnssdInit is delayed quite a lot. The only thing that got it to work was a combination between a timer and capturing the kDnssdPlatformInitialized event.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs.

[stale]

This stale issue has been automatically closed. Thank you for your contributions.

[eme-bennytang]

Hi all,

I think I am seeing this problem here, I found another issue that is related but my code has already implemented that change
#22131

Is there any further update for the OTA state machine for this?