Root node conflict

[iia]

Hi,

In my mesh setup I have 3 nodes which are in MESH_ONLINE mode and connects to a mesh server in my LAN. Where I print the packets received by the server and also send response back to the mesh network. Everything works as expected.

The problem is, if I power up all the three nodes at the same time then there are three root nodes existing for the same mesh network! Even though the nodes are configured with same group ID and SSID prefix. I waited at least an hour but the mesh network did not resolved this situation.

As far as I know there is a mechanism to automatically detect and recover from such root node conflict situation but I guess this mechanism for some reason is not being activated? Is this a known problem?

Can this be a reason that a node was elected to be a root node in the past and this information is being saved somewhere and on each power up it is read and thus the node is becoming root without checking for conflicts? If so where is this information stored and how it can be cleared on each reset?

Any help is much appreciated.

Cheers.

[ESP_Alen]

Hi, we will reproduce the problem you meet.

[ESP_Xutao]

Hi:
I user our demo(https://github.com/espressif/ESP8266_MESH_DEMO)with three ESP-WROOM-02,I reproduce you problem,but it's ok.
Our test environment as follows:
1.Our server(demo_server.py)run on ubuntu.
2.Run three ESP-WROOM-02.
module 1 log(according this log,it's mesh-root):
connected with ASUS_N56U, channel 7
dhcp client start...
ip:192.168.102.150,mask:255.255.255.0,gw:192.168.102.1
Got IP, IP:192.168.102.150Mask:255.255.255.0GW:192.168.102.1
bcn 0
del if1
module 2 log:
connected with MESH_DEMO_1_2734F7, channel 7
dhcp client start...
ip:2.255.255.3,mask:255.255.255.0,gw:2.255.255.1
Got IP, IP:2.255.255.3Mask:255.255.255.0GW:2.255.255.1
bcn 0
del if1
module 3 log:
connected with MESH_DEMO_1_2734F7, channel 7
dhcp client start...
ip:2.255.255.2,mask:255.255.255.0,gw:2.255.255.1
Got IP, IP:2.255.255.2Mask:255.255.255.0GW:2.255.255.1
bcn 0
del if1
This process takes about 40 seconds.
If you still can not，you can provide you test environment,include your demo and your server evironment.

[iia]

Hi,

I still can reproduce the situation where a mesh network has more than one root node and it is not resolved even after an hour. But when this situation has occurred then resetting one of the modules fixes the problem and the node comes as a child node.

Some notes about my configuration.

1. Group ID of the mesh network is, 1A:FE:34:00:00:00. It is same for all the nodes of the mesh network. The last three octaves of the group ID is zero, could that be a problem?

2. The nodes of the mesh network are configured to use WiFi encryption of type, AUTH_WPA_WPA2_PSK.

3. All the nodes enable mesh as, MESH_ONLINE. As found in mesh.h, what are MESH_SOFTAP and MESH_LEAF_LOWPOWER?


Cheers.

[iia]

What I really need is a way to properly trigger a new root node selection among available nodes of a mesh network.

How can this be achieved? Is there an API for this?

Cheers.

[iia]

Seems like I figured out what was causing the behaviour reported. I was calling, wifi_station_set_config_current() before enabling mesh and that was causing the unusual behaviour.

Cheers.