Addon crashes and does not restart via watchdog on Home Assistant Yellow with ezsp #431
Labels
Comments
|
Home Assistant 2022.12.8 Zigbee2MQTT Current version: 1.28.4-1 after reboot the Z2M everything works. |
|
I have the same issue, seems to be after I installed beta of 2023.1. After restoring backup it worked again only occasionally the report comes back, or zigbee2mqtt crashes and the only solution is a reboot of the yellow. When the addon crashed I also lost some devices in z2m and I have to repair them before they become visible again. This happens almost every day and I can't figure out where this problem comes from. |
|
I have my Zigbee coordinator and a Z-wave stick connected to a RPi, whose only purpose in life is to serve those two out over ser2net. If I need to install updates on the RPi, which require a reboot, or if the network link is unavailable for a period of time, between the RPi and the machine running z2m, then even with Watchdog turned on, I get the bellow error, with no subsequent retries and have to manually restart the container. On the other hand, Z-Wave JS UI MQTT behaves as-expected, and resumes normal operation once the network communications are restored.
|
|
Still the same problem. The first start/restart of the addon always fails with the following log: The next start/restart works fine. My serial settings are as follows: Yet the first start/restart ALWAYS fails :( The next one ALWAYS works. |
|
I am also waiting for a solution to this problem |
|
Still not fixed: |
|
Still persists with 1.30.0-1 (running Home Assistant Yellow) One start-up fails: Next one is working again (not much different in the logs, except the error in the first "Uart reseting" I am currently using a workaround where I let Home Assistant check the logs using a command_line sensor which executes: #!/bin/bash
CHECK=$(tail -n 1 /config/zigbee2mqtt/log/current/log.txt | grep "Starting zigbee-herdsman")
if [[ -z $CHECK ]]; then
echo "0"
else
echo "1"
fiand restarts the Addon/Container using an automation. This is actually only needed because the Z2M Addon would sometimes crash (e.g. during an OTA update of some bulbs), then gets restarted by HA and then gets stuck.. |
|
@tienducle Nice workaround! Could you please describe the logic of the automation? "Starting zigbee-herdsman" would be in the log for a long time, so your automation will always trigger until the log is rotated. How do you make it restart only once if it's stuck? |
|
This is the automation: The sensor config is - platform: command_line
name: Z2M Failure
command: /config/checkz2m.sh
value_template: "{{ value == '1' }}"
scan_interval: 30So when the container gets stuck on the "Starting zigbee-herdsman" log message, the sensor would turn from False to True and that triggers the automation which then waits another minute until calling the restart service. When the container restarts, the log file is rotated automatically (not sure if that's a default setting in Z2M but I don't remember changing anything for that) so that after ~30 seconds, the sensor changes to False again. |
|
This issue is stale because it has been open 30 days with no activity. Remove stale label or comment or this will be closed in 7 days |
|
Any updates on this? |
|
This may be fixed in 1.30.2-1. At least now restart works for me. Probably related to the most recent fixes in ezsp adapter. |
|
This issue is stale because it has been open 30 days with no activity. Remove stale label or comment or this will be closed in 7 days |
|
still having this issue. what non-workaround is there for this? is there a way we can see why or how watchdog works to fix it? |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Description of the issue
UI is not available: 502: Bad Gateway.
The log freezes with the below.
Addon version
1.28.4-dev (commit #b0e352b)
Platform
Home Assistant 2022.11.5
Supervisor 2022.11.2
Operating System 9.3
Frontend 20221108.0 - latest
Logs of the issue (if applicable)
No response
The text was updated successfully, but these errors were encountered: