Openthread Border Router cannot access itself to public IP

[bgorostidi]

Hello,

I have built basic_thread_border_router (v1.1 ESP Thread Boarder Router SDK) and flashed on ESP Thread Border Router dev board with Ethernet adapter.
Border Router properly attach to Wifi or Ethernet and create a Thread network. End Devices can join the network and access public IP (ping 8.8.8.8 succeeds) but Border Router cannot itself ping public IP (ping 8.8.8.8 fails).
Web server is working.

Main goal is to add websocket client on BR to be managed by a remote server and ntp capabilitiy.

How can I configure the BR to access Internet?

BR information:

> debug

$ version
openthread-esp32/c8fc5f643b-f32c18bc0; esp32s3;  2025-01-10 15:26:47 UTC
$ state
leader
$ rloc16
ec00
$ extaddr
82e7fdebfabef6fb
$ ipaddr
fd00:db8:a0:0:0:ff:fe00:fc10
fda6:d29f:d2ef:1:f297:4cfc:cd97:1ebf
fd00:db8:a0:0:0:ff:fe00:fc00
fd00:db8:a0:0:0:ff:fe00:ec00
fd00:db8:a0:0:61da:30d8:4cd2:8cce
fe80:0:0:0:80e7:fdeb:fabe:f6fb
$ ipmaddr
ff33:40:fd00:db8:a0:0:0:1
ff32:40:fd00:db8:a0:0:0:1
ff02:0:0:0:0:0:0:2
ff03:0:0:0:0:0:0:2
ff02:0:0:0:0:0:0:1
ff03:0:0:0:0:0:0:1
ff03:0:0:0:0:0:0:fc
$ channel
15
$ panid
0x1234
$ extpanid
dead00beef00cafe
$ netdata show
Prefixes:
fda6:d29f:d2ef:1::/64 paos low ec00
Routes:
fc00::/7 sa med ec00
fda6:d29f:d2ef:2:0:0::/96 sn low ec00
Services:
44970 5d fd000db800a0000061da30d84cd28cced120 s ec00 0
Contexts:
fda6:d29f:d2ef:1::/64 1 c
Commissioning:
16290 - - -
$ netdata show -x
08040b023fa203140040fda6d29fd2ef0001070211400504ec00f1000b1980015d0d14ec00fd000db800a0000061da30d84cd28cced12003080007fc0103ec001003130060fda6d29fd2ef0002000000000103ec00e0
$ partitionid
1856584882
$ leaderdata
Partition ID: 1856584882
Weighting: 65
Data Version: 247
Stable Data Version: 36
Leader Router ID: 59
$ child table
| ID  | RLOC16 | Timeout    | Age        | LQ In | C_VN |R|D|N|Ver|CSL|QMsgCnt|Suprvsn| Extended MAC     |
+-----+--------+------------+------------+-------+------+-+-+-+---+---+-------+-------+------------------+

$ childip
$ neighbor table
| Role | RLOC16 | Age | Avg RSSI | Last RSSI |R|D|N| Extended MAC     | Version |
+------+--------+-----+----------+-----------+-+-+-+------------------+---------+
|   R  | 0x1c00 |  13 |      -26 |       -26 |1|1|1| 7eae16e50f4a3ce6 |       5 |
|   R  | 0x5c00 |  11 |      -19 |       -19 |1|1|1| 7eaa2699ae40b536 |       5 |

$ router table
| ID | RLOC16 | Next Hop | Path Cost | LQ In | LQ Out | Age | Extended MAC     | Link |
+----+--------+----------+-----------+-------+--------+-----+------------------+------+
|  7 | 0x1c00 |       23 |         1 |     3 |      3 |  13 | 7eae16e50f4a3ce6 |    1 |
| 23 | 0x5c00 |        7 |         1 |     3 |      3 |  11 | 7eaa2699ae40b536 |    1 |
| 59 | 0xec00 |       63 |         0 |     0 |      0 |   0 | 0000000000000000 |    0 |

$ nexthop
| ID   |NxtHop| Cost |
+------+------+------+
|    7 |    7 |    1 |
|   23 |   23 |    1 |
|   59 |   59 |    0 |
$ eidcache
fda6:d29f:d2ef:1:5ce4:30dc:6189:4445 1c00 cache canEvict=1
$ deviceprops
PowerSupply      : external
IsBorderRouter   : yes
SupportsCcm      : no
IsUnstable       : no
WeightAdjustment : 0
$ srp client state
Enabled
$ srp client host
name:(null), state:Removed, addrs:[]
$ srp client service
$ srp client server
[fd00:db8:a0:0:61da:30d8:4cd2:8cce]:53536
$ srp server state
running
$ srp server addrmode
unicast
$ srp server host
$ srp server service
$ br state
running
$ br omrprefix
Local: fda6:d29f:d2ef:1::/64
Favored: fda6:d29f:d2ef:1::/64 prf:low
$ br onlinkprefix
Local: fdde:ad00:beef:cafe::/64
Favored: fdde:ad00:beef:cafe::/64
$ br prefixtable
$ br nat64prefix
Local: fda6:d29f:d2ef:2:0:0::/96
Favored: fda6:d29f:d2ef:2:0:0::/96 prf:low
$ bufferinfo
total: 65
free: 55
max-used: 18
6lo send: 0 0 0
6lo reas: 0 0 0
ip6: 0 0 0
mpl: 0 0 0
mle: 0 0 0
coap: 5 10 250
coap secure: 0 0 0
application coap: 0 0 0
Done

[chshu]

@bgorostidi Please refer to the following explanation for details on why ping may fail: Ping Fails.

In general, the ping command from ot_cli fails as expected. For accessing the public internet, it is recommended to use lwIP APIs instead.

[bgorostidi]

Thanks for the explanation.

In fact, I've talked about ping but I've also created a project based on the basic_thread_border_router sample in which I've implemented a sntp service and a websocket client with which I have same problems.
At startup, the device connects to the wifi then starts the SNTP service and the websocket client. Both services fail, whereas if I block the border_router service from starting, the services succeed.

[chshu]

@bgorostidi In the ESP Thread BR v1.1 release, the default netif was set to Thread, causing lwIP to route packets to the Thread interface if no better prefix matched. This configuration has been removed in the master branch. Please apply this commit and try again: commit link.

[bgorostidi]

After several attempts, BR managed to access to public IP. I applied the commit and set the default netif (wifi or eth) once the openthread task is started.

esp_netif_set_default_netif(wifi_netif);

Thank you for taking the time to help me.