UZG-01 running 20240707 crashes (stalls) after 1-2 days

[victorhooi]

Describe the bug

I have flashed XYZ onto a UZG-01 (ZigStar UZG) device. It is running as a Zigbee Coordinator in my network, and is connected to Zigbee2MQTT.

It is being powered by PoE.

However, after approximately 1-2 days, the device appears to disappear off the local network - it is no longer responding to ICMP ping, and the web interface is not accessible either, and Zigbee2MQTT is not able to connect to it.

However, the LAN Ethernet activity lights are still flashing, and the left-most blue activity light is flashing. (I can attach a video if needed - or if the frequency etc is somehow relevant).

Device information

• Screenshot of Zigbee and Device tab from root page of device
• JSON object returned on http://xzg.local/api?action=1&param=all You should remove your credentials first

Unfortunately I can't get the above, as the device's web interface isn't responding.

Is there some way to get debug logs from the device, before I power-cycle or reboot it? Or what else can I retrieve before I try to reflash it again?

To Reproduce Steps to reproduce the behavior:

1. Get a UZG-01.
2. Flash [20240707](https://github.com/xyzroe/XZG/releases/tag/20240707) onto the device.
3. Configure it as a Zigbee coordinator, and point a Zigbee2MQTT instance at it.
4. Wait...profit.

Expected behavior The UZG-01 running XZG firmware should not crash or become unavailable.

Screenshots If applicable, add screenshots to help explain your problem.

Additional context Add any other context about the problem here.

[xyzroe]

I'm also using UZG with 20240707 and don't have such problems. We need to find something special in your setup.

[victorhooi]

Got it - is there some way of retrieving any debug data from the running unit (e.g. via the USB-C port)?

I haven't power-cycled the unit yet, as it's currently in the crashed (wedged) state - I was waiting to check if there might be useful diagnostic information to pull?

[victorhooi]

@xyzroe Am I OK to power-cycle the USG-01? Or is there something I can collect from the box in its crashed state?

And after I power-cycle it - is there something I can setup to collect more verbose logs, or capture diagnostic information on why the unit might be stalling?

[xyzroe]

You're ok. you can build the debug version of firmware it includes more logs.

[victorhooi]

Got it - is there a place I can download the debug build already for 20240707?

Or do you happen to know if this will build easily on macOS? (Otherwise I can see about setting up a Linux VM to do this all on. Sorry if this is a silly question - but should I assume it will cross-compile to ARM fine, from a x64 box?)

Also - how do I enable the debug build? And where will the additional debug information get dumped to afterwards? Or how do you view these logs?

[xyzroe]

I'm using mac os. Just install all dependencies

[xyzroe]

Just select "debug-solo" build variant

[DieMetRik]

I have the same issue after flash firmware > 20240610 Every day near 20:00 my zigbee coordinator rebooted

«My network» > «Keenetic Giga (KN-1011)»
«ZigStarGW» connecting

Keenetic Bot, [07/19/2024 22:11]
«My network» > «Keenetic Giga (KN-1011)»
«ZigStarGW» disconnecting

Keenetic Bot, [07/20/2024 20:01]
«My network» > «Keenetic Giga (KN-1011)»
«ZigStarGW» connecting

Keenetic Bot, [07/20/2024 20:03]
«My network» > «Keenetic Giga (KN-1011)»
«ZigStarGW» disconnecting

Keenetic Bot, [21.07.2024 21:41]
«My network» > «Keenetic Giga (KN-1011)»
«ZigStarGW» connection

Keenetic Bot, [21.07.2024 21:42]
«My network» > «Keenetic Giga (KN-1011)»
«ZigStarGW» disconnection

I back to firmware 20240610 and all OK

My Setup

[MaartenMol]

Any update on this?

[cshuttle]

I am having this issue also, resulting in an unavailable controller several times a day.

[sneak-o-matic]

for me the same

my Device: Zig-Star model: UZG-01

connected via USB/Wifi/LAN

always results into device stalling after a few, max 12h, hours

[airecken]

I am having the same issue.

Device: Zig-Star UZG-01 Hardware: CC2652P7 Firmware: 20240710

Setup as coordinator. Happens with fixed IP or DHCP, powered via USB-C ror POE.

[xyzroe]

Try to disable cron based update availability check

[airecken]

Try to disable cron based update availability check

I updated to 20240914, set frequency to disable. Still automatically reboots itself usually within 15 minutes.

After multiple self reboots the device usually hangs and requires a power cycle.

[MaartenMol]

Any update on this?

[spitfire]

Doesn't happen to me, I'm running a newer version of Zigbee firmware (20240710) though. Have you tried upgrading it?

[shyawnkarim]

This same thing has been happening to me for months. I have one connected via PoE with a static IP and DHCP disabled. Its connected to HA via Z2M and 1 or 2 times a day HA will lose connection with the ZigStar and the zigstar will also become unreachable. I end up unplugging it, plugging it back in, and then it connects and behaves for a bit till i have to restart it again.

[sneak-o-matic]

Now i bumped to latest pre-release (20241001) and then updated zigbee firmware to latest (20240710) and disabled firmware checks. Its running stable since three days. Gonna update if that changes.

edited: corrected zigbee fw version

[airecken]

I installed latest pre-release (20241001) with the newest Zigbee firmware I could find (20240710) and that seems to be stable so far as well. It's definitely lasted longer than before without crashing, so I think that memory leak maybe fixed.

Curious which 20241001 Zigbee firmware you were installing?

[sneak-o-matic]

@airecken i am sorry. that was a mistake by me. you are right: latest zigbee fw is 20240710. i correct that.