I've had some odd operations twice in two weeks with Zigbee devices with my C8 Pro originally running 2.4.2.160.
Stay with me, about a week ago and with no changes to the system my Zigbee devices started not responding/transmitting. they were either intermittent on the communication or very slow or took multiple instances of motion or contact closing/opening to respond. While I foolishly didn't look a closely at the logs then, and having a fair number of Iris v1 devices I ran around the house pulling batteries, disconnecting devices, reconnecting a limited number of devices, rebooting of hub, rebooting of zigbee antenna, rebuilding network and moving of the hub, backing down the firmware and reloading the hub from a backup. Finally to get to the point that over time I was able to reconnect all the devices and given 24 hours everything was communicating well, if not better/quicker than before the craziness began. I chalked it up to interference from other wireless and moved on.
There were a couple last devices that I had not connected back in but was slowly getting to those. Then this morning I started to see where Zigbee devices are acting slow and not responding again. I did add a new sensor last night (ThirdReality tilt) but everything else stayed the same. With a cooler head I checked the logs and am seeing the attached warnings about re-routing. Items like power controllers are needing to be re-freshed (power button pressed on the device) seemingly like when they need to be exercised to find the network.
I have not done anything more this morning to diagnose/fix this problem like reboot or rebuild the network. I have pulled a couple batteries in devices, but I am fighting the urges to go crazy until I hear any suggestions from the community. While I do have 2.4GHz and 5GHz Wifi, I just ran a sweep for interference and let it move a some of the channels around.
Thoughts? In terms of going back further in the logs on warnings, when I click on show more, it doesn't populate. I was hoping that I would see when this problem started and it correlating to when things stopped playing nicely.
It depends on the situation, but for
Motion - 3rd Hall Stair - Iris v1 motion Detector Custom
Motion - 1dt Hall Stair - Iris v1 motion Detector Custom
Contact - Coat Closet Sensor - Iris v1 Contact Sensor - Custom
Motion - 2nd Stair Lower - Iris v1 motion Detector Custom
These drivers have been set for months. I have recently updated drivers/apps in HPM where these drivers came from, but this was between the two rounds of issues and I don't believe these were part of that.
Are these Iris v1 custom drivers publicly available on GitHub (or via HPM)?
Its difficult to guess what these warning messages mean without looking at the code, but if the number 13 in the "0013 Re-routeing" is sent as a warning when ZDO command 0x0013 is received, it really means that these devices are constantly trying to rejoin the Zigbee network. The reasons could be device reboots or Zigbee network congestion / interference (the devices sending ZDO command 0x13 have lost connection).
Many posts in the forum discuss the incompatibility of Iris V1 Zigbee devices with Hubitat hubs. However, I don’t own any Iris devices, so I lack practical experience.
Drivers are publicly available via HPM. I can change, but would expect that if they were a problem, I would have seen it by now. I will put that into my list of things to try next. I went to these drivers from the built in as they provide better battery reporting. And yes, I know I am walking a tight rope on the v1 devices. Since I did the transition using the Hubitat migration tool, I can still use them. Once I lose them I need to change them or send those devices to my C5. Honestly, I was using the C5 as a hot spare/programming tool but have started to decide to actively move more of these v1's to the C5. Unfortunately, the one I started with is being a pain to re-pair.
Another interesting piece of information is that I forgot that I had a reboot schedule running. So the hub has been up for 15 hours now, and would explain why my logs won't go further back. might explain why this problem has renewed.
Also, for those keeping score, I had backed down and am still running 2.4.2.158.
I looked at the drivers code, they are really good! There is no need to change the drivers; the built-in drivers are likely ignoring the re-join command without issuing a warning in the logs, so the problem persists.
In my personal opinion and experience, automatic attempts to recover the connection from the coordinator side (the HE hub) by pinging battery-powered devices from the driver are generally unsuccessful, particularly with older generation Aqara sensors. However, I cannot speak with certainty about Iris V1 devices.These older Iris devices seems to have a lot of 'specifics' too..
As far as I know, Iris v1 devices have no problems with C7/C5 hubs, so moving these back to your old hub will most probably solve the problem.
Allow me to provide a bit of an update. Since my last posts I was limping along with good days and bad days with errors and sporadic communications on my zigbee devices (both newer and Iris v1) for another week thinking that I was close to solving (wasn't really) while waiting for replacement contact sensors to arrive. Last weekend they did and I feverishly swapped them in and spread out sensors between my C5 an C8 Pro hub as I have thought to do for a while.
New sensors on the C5 were working great (responsive) and I was no longer getting error messages in logs for the C8 Pro, but still sporadic communication issues with devices on that hub. Was starting to suppose there was a problem with the zigbee side of my C8 Pro and was gearing up to move them all to the C5.
Starting with a (new) sensor that was in line of sight of the C8 hub that was acting poorly, I performed reset the contact sensor and went to re-pair it to confirm it would recognize it if I hadn't removed it from the hub. During the process of re-pairing not only did that device show that it was already connected to the hub, a mix of existing devices and new contacts showed up as already being connected. Spent the rest of the evening poking at other zigbee devices to check their operation and re-pairing them as necessary with others joining the fun. Since that exercise my issues are essentially gone and I am back to a stable and quick responding mesh.
I have since been making small adjustments and updates in the evenings to settle some of the items that I did in a crash cart mentality. 95% of the Iris v1 items have been removed from the C8 but I don't believe those were the main culprit, though it gave me the push to get them moved off.
Interestingly enough, I have gone and tried to re-pair the Iris V1 sensors to the C5 hub and I have had about a 5% success rate in getting them connected. While I don't really need them anymore and have squeezed more usefulness out of them than I should have with the help of the programmers at Hubitat, I still would love to sprinkle them around the house for less critical tasks.
Thanks to @kkossev for providing a sounding board while I was pulling my hair out and wanting to set something ablaze...
There is no need to change the drivers; the built-in drivers are likely ignoring the re-join command without issuing a warning in the logs, so the problem persists.
