Skyconnect/ZBT-1 recent issues

[real owl]

So I've had recent issues with trel turning off and the border router vanishing from thread networks.

Saw weird stuff in logs too I tried to limit to warnings and above to avoid spam but here's my log

Latest HomeAssistant versions on RPI

Logs are blank in file so I'll copy and paste issues

00:00:00.096 [W] P-Netif-------: Failed to process request#2: No such process

Default: mDNSPlatformSendUDP got error 101 (Network is unreachable) sending packet to 224.0.0.251 on interface
Default: mDNSPlatformSendUDP got error 99 (Cannot assign requested address) sending packet to ff02::fb on interface fe80::2033:9fff:fe16:8c08/veth07441ec/4127

'Trel will not start' is another thing that pops up something about spindle error

[real owl]

Failed to register service Home Assistant OpenThread Border Router #B519._meshcop._udp: Service Not Running
[WARN]-BA------: Result of publish meshcop service Home Assistant OpenThread Border Router #B519._meshcop._udp.local: Invalid state
[ERR ]-MDNS----: Failed to register service 22043b337124b519._trel._udp: Service Not Running
[ERR ]-TrelDns-: Failed to publish TREL service: Invalid state. TREL won't be working.
[12:02:50] INFO: mDNS ended with exit code 256 (signal 6)...

[long jungle]

Is this HAOS? That's pretty odd, looks like OTBR is trying to ask Avahi to publish mDNS records but it's not working for some reason.

Or does OTBR try to do mDNS itself?

[real owl]

HAOS yeah

[INFO]-WEB-----: Running 0.3.0-b041fa52-dirty
listenAddr not specified, using default ::
[INFO]-WEB-----: Border router web started on wpan0
[00:02:03] INFO: Enabling NAT64.
00:00:00.347 [W] P-Netif-------: Failed to process request#12: No such process
Done
Done
Done

00:00:00.799 [W] DuaManager----: Failed to perform next registration: NotFound
00:00:00.837 [W] Mle-----------: Failed to process Link Accept: Security
s6-rc: info: service legacy-services successfully started
00:00:00.843 [W] Mle-----------: Failed to process Link Accept: Security
00:00:01.099 [W] Mle-----------: Failed to process Link Accept: Security
00:00:01.108 [W] Mle-----------: Failed to process Link Accept: Security

Trel stops working after a while and then like a day or two after that the border router leaves the network entirely until I restart the add-on.

Even if the add-on is still running (altho I set watchdog to restart it if it crashes)

[real owl]

It used to be fine, I'm not sure why it's doing this nowadays. I don't remember setting avahi up myself, nor do I remember changing mdns settings

Also is the sky connect Firmware flasher permanently down? Was gonna try using that to see if it helps

Finally got logs working

Even bigger one

[real owl]

Perhaps this part might help answer?

00:01:24.581 [W] DuaManager----: Failed to perform next registration: NotFound

Default: requestIOEvents called with fd 1024 > FD_SETSIZE 1024.
mdnsd: mDNSPosix.c:2527: requestIOEvents: Assertion `0' failed.
[WARN]-MDNS----: DNSServiceProcessResult failed: Service Not Running (serviceRef = 0x55bc12a3b0)
[WARN]-MDNS----: Need to reconnect to mdnsd
[ERR ]-MDNS----: Failed to register service Home Assistant OpenThread Border Router #B519._meshcop._udp: Service Not Running
[WARN]-BA------: Result of publish meshcop service Home Assistant OpenThread Border Router #B519._meshcop._udp.local: Invalid state
[ERR ]-MDNS----: Failed to register service 22043b337124b519._trel._udp: Service Not Running
[ERR ]-TrelDns-: Failed to publish TREL service: Invalid state. TREL won't be working.
[12:38:44] INFO: mDNS ended with exit code 256 (signal 6)...
[12:38:45] INFO: Starting mDNS Responder...
Default: mDNSResponder (Engineering Build) (Dec 3 2024 17:53:13) starting
Default: mDNSPlatformSendUDP got error 99 (Cannot assign requested address) sending packet to ff02::fb on interface fe80::c864:83ff:fe7e:d13b/veth7f93f77/8713
Default: mDNSPlatformSendUDP got error 99

[real owl]

1d.03:22:12.946 [N] MeshForwarder-: Failed to send IPv6 UDP msg, len:94, chksum:7ae0, ecn:no, to:8ab35ea07d62db38, sec:no, error:Abort, prio:net, radio:trel
1d.03:22:12.946 [N] MeshForwarder-: src:[fe80:0:0:0:5861:f22f:ccdd:dc30]:19788
1d.03:22:12.946 [N] MeshForwarder-: dst:[fe80:0:0:0:88b3:5ea0:7d62:db38]:19788
1d.03:45:26.660 [N] MeshForwarder-: Failed to send IPv6 UDP msg, len:64, chksum:c6e2, ecn:no, to:0xb000, sec:yes, error:Abort, prio:low, radio:trel
1d.03:45:26.660 [N] MeshForwarder-: src:[fd6e:d157:2b4:cdbf:0:ff:fe00:5000]:61631

[long jungle]

interesting - that says radio:trel, which means that it's trying to send the message by relaying it through another thread border router somewhere else which it discovered via mdns

which means that the mDNS setup did work, and it did discover another border router via mDNS for trel.

[real owl]

It shows up in trel for a short while like maybe a day before trel stops working on HomeAssistant.

And then only the other trel hubs show up with homeassistant being listed under meshcop. Sometimes it even drops out of that. With 'openthread' integration showing error until I restart add-on, I'm assuming because it crashes.

I think something is wrong with the firmware since I can’t seem to set it up again via skyconnect box in system settings. And ofc the webflasher is still down.

[real owl]

Any way to fix it that you see here?

Not sure what ‘unexpected routers’ are though…

[real owl]

Will prolly be my last log report till I get a request for more, to avoid spamming but this is one of the last errors before it just crashes out of my network entirely and has 0% usage in the add-on stats page. It didn't even always crash outright before needing a restart.
(Ofc I mean Before I attempted to set it up for auto-reset, with watchdog)

dst:[fe80:0:0:0:88b3:5ea0:7d62:db38]:19788 1d.01:39:56.832 [N] MeshForwarder-: Failed to send IPv6 UDP msg, len:64, chksum:c75f, ecn:no, to:0xb000, sec:yes, error:Abort, prio:low, radio:trel
... Failed to get forwarded frame priority, error:NotFound, len:35, src:0x8800, dst:0x9c00, sec:yes
1d.02:05:09.977 [N] MeshForwarder-: Failed to send IPv6 UDP msg, len:94, chksum:164c, ecn:no, to:8ab35ea07d62db38, sec:no, error:Abort, prio:net, radio:trel

---

"1d.02:16:55.801 [N] MeshForwarder-: Failed to get forwarded frame priority, error:NotFound, len:35, src:0x8800, dst:0x9c00, sec:yes
1d.02:16:55.924 [N] MeshForwarder-: Failed to get forwarded frame priority, error:NotFound, len:35, src:0x8800, dst:0x9c00, sec:yes

1d.02:35:17.641 [N] MeshForwarder-: Failed to send IPv6 UDP msg, len:94, chksum:c629, ecn:no, to:8ab35ea07d62db38, sec:no, error:Abort, prio:net"
Then something about noACK & DuplicatedFrames or 'Reassembly error'
"07:37:43.473 [N] MeshForwarder-: Dropping rx frag frame, error:Duplicated, len:50, src:0x2800, dst:0xffff, sec:yes, tag:35037, offset:0, dglen:86"
"Default: mDNSPlatformSendUDP got error 88 (Socket operation on non-socket) sending packet to ff02::fb on interface <<ERROR: %#a used with unsupported type: 161>>/<<NULL>>/-2143420412"

---

Also Tried following any suggestions I could find for similar issues (including but not limited to) stuff like:

• Checking my internet network,

• Changing Baudrate & Turning off automatic firmware flashing before I reset the border router to defaults.

• Restarting HomeAssistant, from app after updates.

• Opting into betas to see if fresh update fixes it.

• Switching to different storage to see if it's an issue with SD Card or SSD. (I think I even moved to another device entirely at one point)

• Powering off RPi-5 by unplugging and replugging,
  Then repeating all I previously did again in multiple different orders/combos to see if some critical errors change.
  Heck, I even tried flashing to ZigBee before going back to (Open)Thread afterwards to check the logs on that process, but nevertheless it still proceeds to show weird errors in the log after all the same reports like the ones I referenced to here.
  All seems to pop up before I get the initial issue with TREL saying it won't start in the log, and then shutting down/not appearing in my network service browser,
  Which is followed by the 'SkyConnect-ZBT-1' Border Router no longer being shown in _meshcop and then disconnecting from my thread network. The add-on ends with 0% usage and ZBT border-router no longer showing up in any thread network apps. ||(Including being missing from the HASS Network list on the apps ''Thread'' Integration's config page.)||

• Wasn't able try newer versions for my 'Connect-ZBT' (Nor do I have another to try, so this is still just my same old SkyConnect-Branded stick that I've had for a while) BC there is no OpenThread beta firmware file listed on your flasher site that I can try in-order to see if the bugs are tied to any change in the most recent SL Gecko/firmware update.
  I would sincerely appreciate it if someone could give me a new beta update file to test with the webflasher so I can share the logs from that here aswell.
  (Saw some interesting stuff when trying to reinstall the same firmware that's already on the stick)

[real owl]

Even more new errors now

Default: mDNSPlatformSendUDP got error 88 (Socket operation on non-socket) sending packet to ff02::fb on interface fe80::2c59:29ff:fef7:b8f7/<<NULL>>/0
Default: mDNSPlatformSendUDP got error 88 (Socket operation on non-socket) sending packet to ff02::fb on interface fe80::2c59:29ff:fef7:b8f7/<<NULL>>/0
Default: mDNSPlatformSendUDP got error 101 (Network is unreachable) sending packet to 224.0.0.251 on interface fd2b:20d4:672:30cd:b6e7:7588:372c:f1c8/wlan0/3
[WARN]-MDNS----: DNSServiceProcessResult failed: Service Not Running (serviceRef = 0x558fc8c230)
[WARN]-MDNS----: Need to reconnect to mdnsd
[ERR ]-MDNS----: Failed to register service Home Assistant OpenThread Border Router #DC30._meshcop._udp: Service Not Running
[WARN]-BA------: Result of publish meshcop service Home Assistant OpenThread Border Router #DC30._meshcop._udp.local: Invalid state
[ERR ]-MDNS----: Failed to register service 5a61f22fccdddc30._trel._udp: Service Not Running
[ERR ]-TrelDns-: Failed to publish TREL service: Invalid state. TREL won't be working.
[02:34:41] INFO: mDNS ended with exit code 256 (signal 11)...
Default: mDNSPlatformSendUDP got error 88 (Socket operation on non-socket) sending packet to ff02::fb on interface <<UNSPECIFIED IP ADDRESS>>/<<NULL>>/-2143420412

[real owl]

And tried switching to dev build but

Default: mDNS_Execute: SendResponses didn't send all its responses; will try again in one second
Default: mDNS_Execute: SendResponses didn't send all its responses; will try again in one second
Default: mDNS_Execute: SendResponses didn't send all its responses; will try again in one second
00:00:16.158 [W] DuaManager----: Failed to perform next registration: NotFound
Default: mDNSPlatformSendUDP got error 99 (Cannot assign requested address)

[compact helm]

i am having similar issue

have you fixed this yet?

[long jungle]

@compact helm pretty sure your issues are unrelated.

@real owl your HA system is on wifi? at least some of the errors being reported appear to issues with wifi communication rather than thread. Any chance you could connect the device to an ethernet cable to see if that stabilizes it?

[compact helm]

It’s on Ethernet.

[long jungle]

@compact helm you seem to have a different issue; to avoid confusion it would be best to discuss your issue in a separate thread.

[compact helm]

Will do.

[real owl]

Ethernet yeah

Altho mine seems to be two issues, one was resolved by yeeting my addons (I think my Pi 5 8gb ram couldn't handle all of them running)

But made Github issue and there does some to be weird results in some of the logs that they're looking into https://github.com/home-assistant/core/issues/134047

Hopefully there's an easier way to check the 'infra' backbone issue they mentioned tho

[long jungle]

can you check to make sure that your wifi is completely disabled? you don't want it and ethernet at the same time

if an application starts up when wifi happens to be connected, and uses that interface, then the wifi disconnects, you could see this sort of issue happen even if the system still have a stable connection via ethernet