I am having some intermittent issues with one of my C7s. Over the last few weeks, at different times, I have noticed laggy performance or non-performance when rules are requested to be executed. This is across a number of devices and a number of rules (when motion is detected - turn on light, when scene controller button is pushed - turn on fan, when bathroom light 1 is turned off – turn off bathroom light 2 and bathroom light 3 etc.).
In addition to the sluggishness, I have had to reset the hub several times due to complete non-responsiveness.
I have not added a new device to the network
I have not removed a device from the network
Nothing has been changed related to the hardware
I have updated the Hub to the latest software (2.3.6.146)
o NOTE: The issues existed prior to the software update
I’m not really sure what the next best step for troubleshooting would be.
I’ve had similar symptoms that a reboot would resolve. Just needed to do the reboot before it locked up.
I would see the memory getting low after a month or so online. In the last week or two it gets critical in a few days. Currently it is under 200MB with 2 days, 3 hours of uptime.
I haven’t changed anything (devices, apps, or firmware) in a while - since the last update.
Please check this post below.
I would follow the instructions for doing a local backup and restore (which also does a soft reset as part of the restore). This has been shown to fix a rapid memory decline issue that popped up in 2.3.6 firmware. Low memory would cause what you are both describing.
Or, maybe they could just address the problem that popped up in 2.3.6 firmware. I THINK I have my system stable now, in part after doing a restore on my hubs, as you describe. But 2.3.6 was not a smooth upgrade. Hoping the team does a post mortem to see what might be done differently for the next major firmware update. This one, unlike any I have experienced before, was painful.
There is a beta that has been in the works for a while, with one focus being the memory decline over time. So hopefully it will fix whatever happened with 2.3.6
Overall, I've been very happy with my Hubitat devices and I appreciate that every once in a while, things don't go as planned. Hopefully this beta will prove to be the fix here. Fingers crossed - thanks for sharing this info.
Thank you for the replies. If this is a known issue with 2.3.6, I am happy to wait for a release that provides resolution.
I have ~150 Z-Wave devices on my network. A little over a year ago, I suffered through months of intermittent issues, eerily similar to the issues I am currently experiencing. After many (many, many) hours of troubleshooting, I bit the bullet and purchased four additional C7s. I spread the z-wave devices and rules out across five C7s that are strategically placed around my home. This immediately solved every issue that I had been struggling with. For greater than a year, my network worked (mostly) flawlessly.
Two of my five C7s are running 2.3.6 software. One of the two C7s running 2.3.6 is having issues. I did not upgrade to 2.3.6 until AFTER I had non-responsiveness issues. I have the “Rebooter” app that runs once a week on all five of my C7s. Not sure if that is a good thing or bad thing, but it was born from necessity as it resolved intermittent memory issues.
As you can see from the attached screen grab of the “freeOSMemoryHistory”, I am experiencing an almost predicably linear increase in memory utilization. Even when memory availability is not critically low, I am still having the issues.
My fear (and motivation to post) was that I had a hardware issue. However, it sounds as if it could be software and that a fix is imminent. At this point, my issues are annoying, not unbearable. I will plan on patiently awaiting an updated release and will assume that it will address my specific problems.
Thank you again to everyone that took the time to read and respond.
That decline on your memory is perfectly normal with the JVM OS that is running. It just keeps caching things which slowly appears to eat up memory. The memory is used by cache but can be freed back up by the OS as needed.
The issue that came up in 2.3.6 you would see a hub go from a fresh reboot to below 100mb free in 2-3 days, and eventually the hub UI would not be reachable anymore forcing a hard reboot.
The fix for that is simple, download a backup and then restore it. This cleans up the database in the process, which for some reason fixes the issue. I am not privy to why this is the case. Most people do not see it return again once fixed.
I rolled my software back to 2.3.4. Ran a backup. Performed a soft reset. Restored with the backup file. Within a few hours, the Hub Mesh for all devices was "Offline" and the web UI was unreachable. However, all of the Rules still operated as expected (with the exception of the rules leverage devices paired to other C7s via Hub Mesh) - without delay or sluggishness.
Power cycling brought everything back up.
All of my C7s have statically assigned IP addresses. In the interest of eliminating a possibility that the problem is network related, I changed physical connection to the network. Any other thoughts on next steps?
This typically means that the hub is still functioning fine but it got booted off the network and could not reconnect. The reconnect logic was actually improved in one of the more recent firmware updates.
Are they connected directly to the main router or through other switches, access points, or bridges?
Is the static IP set on the hub, or is it a DHCP reservation on the router? (The DHCP reservation is recommended).
The static IP is set by way of a DHCP reservation on the router. The hub connects to the network via an ethernet connection to a 48 port switch. The switch connects to the router via SPF+.
I have not yet turned on the "Auto Reconnect" feature.
That setting is actually removed in the newest firmware. It was improved and it set at the optimal setting and cannot be changed. If it seems like you have ruled out the new firmware from causing the issue, you may want to update to current again.
Also, if it goes offline again check the diagnostic tool on port 8081, if that works then the Hub web UI crashed, if it does not work the network interface crashed or the hub got knocked offline.
Do you have anything on your LAN with jumbo frames enabled? That is known to knock the hub offline, requiring a reboot to fix. A lot of people turn this on for a NAS and it can cause more problems than it is worth. The speed boost is negligible in a home environment.
