Unexpected NWK Leave Request

Hi Andrew,

Can you please share the packer sniffer log (in psd format)?

Thanks,

OD

Hi OD,

Thanks for your help. I have attached the psd file.

1602.drop-out-01.psd

Andy

Hi,

I have uploaded a new packet sniffer file.

My set up is very simple: one ZC, four ZED's. Each ZED transmits every minute two bytes of data and goes back to sleep. The packet sniffer log shows everything working well for 32 minutes. Then at message number 644 things start to go wrong.

Message 649: Leave Request

The ZED sends a message as normal but it does not receive an ACK from ZC after it sends its Data Request. So it resends a Data Request 2ms later to which it does get an ACK. However, the ZC has decided that it wants the ZED to leave and issues a Leave Request.

One down, three to go.

Message 1976: Leave Request

Some 82 minutes later, another ZED is told to leave, but this time there was no missed ACK. See message 1972 which is the ZED sending its two bytes of data, it gets an ACK, sends its Data Request, gets an ACK, but then is issued a Leave Request.

At the time of writing the remaining two ZED's are still working and on the network. All ZED's have the same firmware and are identical in hardware.

My next job is to repeat this and try it with ZIGBEE_CHILD_AGING disabled.

Is there a way to determine from the packet sniffer why a Leave Req was issued?

Many thanks,

Andy

Update: disabling ZIGBEE_CHILD_AGING makes no difference.

I have discovered the possibility that the first ZED to join never leaves the network, only those which are joined afterwards.

Andy

update: they all leave the network after some time has elapsed. I left it overnight and by the morning all ZED's have been issued a Leave Request.

Hi  Andrew Bakin

        do  you  find a way to slove this issue?

      i  have encountered  in HA1.2.2a 。

       it  just  let the ZED link the ZC ，and  do  nothing。

       sometimes  i saw the ZED sending  the rejoin request and then the ZC sends a Leave CMD just for it。

      i using  the POLL RATE mode of the Child Aging

       do  you know why does ZC will  auto sends the leave CMD？

BR?

I am facing exactly the same issue at present, when using the ZHA-1.2.0 release.

We got 1 coordinator/several routers and 50+ endevices, everything works fine at the beginning. But after running for a long time, some endevice become offline.

As we have enabled a 7x24 packet capturing, it is easy to find the root-cause. When this happens, the coordinator or router would send out a NWK leave packet to the endevice, with re-join flag set to True. But the endevice would not start rejoin procedure but still keeps sending Data Request.(The endevice's behavor is not stable, sometimes it could rejoin sometimes it would not...)

I am also expecing to know the reason on these "Unexpected NWK Leave Request"[1]

Hope Ti enginners' Input...

Thank you

[1]packet captured example:

Endevice: 0x0b61

Router: 0xf1a5