JN 5169 Tables was cleared

[weisheng]

There are 30 router in the network, and has work fine 2days,and then an device(A)'s MAC was cleared from Address Map Table,Route Table and Neighbour Table.So when A report his attribute to coordinator, we can not determine who is reporting(The Corresponding Relation between MAC addr and Short Addresses on Coordinator was missed)

1.My envirorment:

1) SDK：JN-SW-4170 v1840
2）Coordinator：Base on JN-AN-1216-Zigbee-3-0-IoT-ControlBridge-v1011, we modify code about Uart1,RF signal handle,table size, etc.

3）30 Routers joined the network

2.Uart print some suspicious information: network status 12(sNwkStatusIndicationEvent.u8Status), no other Error logs.

3.The field ‘auth’（uAncAttrs.bfBitfields.u1Authenticated）change to 0

4.Sniffer log attach below.

[mario_castaneda]

Hi Wei,

I am checking this issue.

Did you print the binding table? Do you have the terminal log?

vDisplayBindingTable(),  appZpsExtendedDebug.c

The table is saving data in order, so did you see or know the specific device that was erased from the table?

I mean, is it the first o the last device that joined to the Zigbee Network?

Regards,

Mario

[weisheng]

Hi Mario,

Sorry to reply so late.

I print the "Neighbour Table","Address Map Table","Routing Table" as log attached below, see device 0012000012345613 (zigbee test tool),  it's Map table and mac table was missed，the log is "[2019-08-07 16:21:33 646] vEndPointWriteEvents MacAddr is 0, srcAddrMode 2, shortAddr d50a".

The Device is neither the first nor the last device that joined to the Zigbee Network.

I doubt that at the time "[2019-08-07 16:20:26 333] Nwk Status: Addr:d50a Status:12" the Map table and mac table was cleared, but I can not sure.

Note that in the terminal log, when printed "[2019-08-07 16:21:18 695] [ERROR]Purge Short:d50a Routing Table", the device 0012000012345613 had already been cleared from "Address Map Table"  and "Mac Table", so we clear it from route table.

[mario_castaneda]

Hi Wei,

I was checking your sniffer log and the terminal log.

Why is the coordinator sending the permit to join disabled? 

Could you please read out the EEPROM memory using the production flash programmer?

https://www.nxp.com/docs/en/user-guide/JN-UG-3099.pdf 

Regards,

Mario

[weisheng]

Hi Mario,

   We want to form a start topology network，so we sending the permit to join disabled.

    Sorry，the problem environment is gone，I can not read out the EEPROM memory.It will take several days to recreate this problem.

Regards,

Wei

[mario_castaneda]

Hi Wei,

Please let me know if you still have this issue.

Regards,

Mario

[weisheng]

Hi Mario,

I still have this issue,please help me out,any suggestion is good for me.

Regards,

Wei

发自我的iPhone

[mario_castaneda]

Hi Wei,

Could you please send a ZPS_eAplZdpMgmtLqiRequest moving the index?

The coordinator is asking for the devices that it has in the routing table.

Please let me know your findings.

Regards,

Mario

[weisheng]

Hi Mario，

      I send  ZPS_eAplZdpMgmtLqiRequest and the log was attach below.

      In the log, 0x00124B001BE0EDB2 is real route device, 0012000012345601~0012000012345605  is virtual device(our zigbee test tool),and they did not response to ZPS_eAplZdpMgmtLqiRequest.

      I can't find useful information here，do you find any useful information?

      Look forward to your reply.

Regards,

Wei

[mario_castaneda]

Hi Wei,

Just to confirm, the tables are cleared after 2 days.

The ZPS_eAplZdpMgmtLqiRequest is for the Coordinator. it will look at the list of neighboring nodes. You won't see any command in the air. I want to be sure if the coordinator is trying to read the same address for the devices that I thas saved.

Regards,

Mario

[weisheng]

Hi Mario,

   Some tims the tables are cleared after 2 days, and some times it was cleared just a few hours later.

   I am testing ZPS_eAplZdpMgmtLqiRequest.

Regards,

   Wei

[mario_castaneda]

Hi Wei,

Are you remove or sending leave commands to your devices?

Regards,

Mario

[weisheng]

Hi Mario,

    When we want to remove the end device,we did send leave command to devices,but when the problem occurred,we did not send the command.

Regards,

    Wei

发自我的iPhone

[mario_castaneda]

Hi Wei,

Yes, I want to confirm if the routers are sending the device update to the coordinator.

Does the router, in this case, leave the network? or Trust Center sends the leave command.

Regards,

Mario

[weisheng]

Hi Mario

   The routers didn't send the device update to coordinator.

   The router didn't leave the network and Trust Center didn't sends the leave command,in this case.

Regards,

Wei

[mario_castaneda]

Hi Wei,

Could you please confirm your table size values?

The MAC Address table should be sized according to the combined sizes of the Neighbour table and Address Map table

if you have 30 devices in your network, your set up should be like this.

Neighbour Table 26

Routing Table 30

Address Map 36

Mac Table 62

https://www.nxp.com/docs/en/user-guide/JN-UG-3113.pdf B.5.3 MAC Address Table

Regards,

Mario

[weisheng]

Hi Mario,

   I am checking  the table size values,try to modify the Mac table size to 62,and then see if the problem still exists.But I think the neighbor table should be set to 30. Why 26?

   Thank you for your advice and look forward to your reply.

Regards,

Wei

[mario_castaneda]

Hi Wei,

The MAC Address table should be sized according to the combined sizes of the Neighbour table and Address Map table. 26+36 = 62

https://www.nxp.com/docs/en/user-guide/JN-UG-3113.pdf B.5.3 MAC Address Table

Regards,

Mario