I recently had my Hubitat hub lockup while I was away. It required a manual power cycle when I returned to bring it back to normal operation. It there a way to set a watchdog timer that automatically restarts the hub?
If the hub itself has stopped functioning, then you will likely need to setup something outside of Hubitat to monitor it's availability in some way, such as a ping request and/or HTTP response. You could look at setting up something in Node RED or similar, if you have a spare always-on machine where you could install it. That said, monitoring is one thing, but I would be cautious about automating the power-cycling of the HE hub. If the system that is monitoring HE's availability get's it wrong (there could be other reasons for the monitoring system not being able to communicate with the HE hub) and power cycles HE, that can produce database corruption or other issues.
So while I don't see a problem with setting up monitoring and alerting of some kind, as for automating a response, I would suggest looking into why your hub became / becomes unresponsive, and try to resolve the underlying issue, rather than rebooting automatically.
1 Like
Thanks for the fast reply. Since the hub had locked up and was non-responsive, there wasn't any obvious real-time way to determine why. Are there device logs that are accessible that might be instructive in determining when or what the hub was doing at the time it hung? Is there a live debug capability on the hub?
The following is what I found in the 'past logs':
dev:652023-04-07 04:14:50.373 PMinfoBasicReport value: 0 ok
dev:652023-04-07 04:14:50.356 PMdebugparse description: zw device: 14, command: 2003, payload: 00 , isMulticast: false
dev:652023-04-07 04:14:50.260 PMdebugrefresh()
dev:652023-04-07 04:14:49.980 PMwarnconfigure...
sys:12023-04-07 04:14:40.539 PMinfoZ-Wave Discovery Running
sys:12022-11-25 05:35:16.787 PMinfoFinished Z-Wave Network Repair
sys:12022-11-25 05:35:10.770 PMtraceZ-Wave Node 10: Repair is done.
sys:12022-11-25 05:35:04.783 PMtraceZ-Wave Node 10: Repair is running
sys:12022-11-25 05:35:02.280 PMtraceZ-Wave Node 10: Repair is pinging the node
sys:12022-11-25 05:35:02.279 PMtraceZ-Wave Node 10: Repair starting
sys:12022-11-25 05:35:00.277 PMinfoRetrying Failed Nodes
sys:12022-11-25 05:34:54.435 PMtraceZ-Wave Node 07: Repair is done.
sys:12022-11-25 05:34:51.274 PMtraceZ-Wave Node 07: Repair is running
sys:12022-11-25 05:34:49.036 PMtraceZ-Wave Node 07: Repair is pinging the node
There isn't any log entries from 11/25/2022 till 4/7/2023 but the hub appeared to be operating (switches turning an/off) according to scheduled tasks until mid March 2023.
That does seem odd.... Unlikely, but not impossible. The creation of logs is dependent on the devices / apps and the settings for whether to produce log entries in different circumstances. It may be worth upgrading to the latest firmware version for the HE hub and monitor things some more, unless others / HE support can identify any other possible causes.
If it continues to happen the HE developers can access additional logs that are produced for their internal troubleshooting, so it may be worth reaching out to them if needed.
My hubitat just crashed now, too. It had been running for about a year just fine. I had just installed the latest Z-wave firmware and it crashed about a week or two after that. I'm not sure if it's related. I do make extensive use of MakeAPI. I can ping it, but if I try to connect to port 80 its says connection refused.
How do we reach out to the HE devs? This is critical and I may have to start looking for more reliable solutions. I was hoping I would not have to debug the HE software. 
when it crashed can you get to yourhubip:8081?
Have there been any changes on your network? Something such as enabling jumbo frames?
Are you seeing any devices spamming the logs?
1 Like
Nothing has changed on the network. Cool, yes, I can get to 8081. I entered the MAC address. But I don't see any way to access any logs in the Diagnostic Tool menu.
I looked at the logs after rebooting and did not see any spamming. The last snippet before it crashed is this:
[dev:74]2023-04-08 06:00:06.517 AM[info]madonna was turned on
[dev:74]2023-04-08 06:00:06.513 AM[debug]parse:zw device: 1A, command: 2503, payload: FF FF 00 , isMulticast: false
[dev:74]2023-04-08 06:00:06.504 AM[info]madonna was turned on
[dev:74]2023-04-08 06:00:06.501 AM[debug]parse:zw device: 1A, command: 2003, payload: FF , isMulticast: false
[dev:76]2023-04-08 06:00:03.728 AM[info]borat was turned on [digital]
[dev:74]2023-04-08 06:00:03.157 AM[debug]on()
You can't... That said, do a restore to yesterday's backup (this is essentially a soft reset and will ensure there is no corruption in your database).
For you and the OP, fill out a warranty claim:
Since support has already been tagged, I'm sure they'll reach out in the next day or two. With your hub ID, they should be able to check the engineering logs to see if anything sticks out.
1 Like
This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.
