As the title says I've been having issues with the zigbee radio reboots. C-8 on release .200. This has been going on through numerous releases. I've spent the last couple of weeks trouble shooting and have come up empty. I thought it was a bad device but took it out and the issue persists. I've been through numerous channel changes from 15 to 26, yes, the low power 26 as I hadn't tried it yet, along with power changes. Some channels are worse than others but all result in zigbee radio reboots. On 3 occasions the hub came back up without a zigbee radio. Shutdown and power pull brought it back each time. Tagging @gopher.ny and @bobbyD for someone to please take a look at the engineering logs in hopes of finding a solution or at least a cause. Or anyone else @support-agent@support_team who can help. I will PM the hub id and whatever else is needed when the lucky one gets the case. Thank you for your time.
Edit: Forgot to mention, checked router channel, changed that as well, moved it. It's about 8 feet away w/ no line of site.
Thanks. That's a good question and another thing I forgot to mention. I only have 5 Z-Wave devices and they appear to be humming along. I watched/checked the logs and nothing stood out. Hubs been up about 18 hours and I've only had 2 radio reboots. Nothing looks bad to me. Admittedly, I'm not sure I'd know if it was but I've seen it a few times over the years now.
But if you notice .. its always for a very short time.
But I have found is alot of devices do not like weird zigbee channels.
20 seems to be the Sweet spot.
I have seral devices that refuse to work or pair on any chan but 20 < for what ever its worth.
also if you set you wifi to chan 1 and set zigbee to 20 8-12 power.
you may have better results.
I also suffer with the zigbee radio going offline briefly, some days more than others. I am using channel 20 and I have zigbee only devices. I will try cranking up the zigbee power, currently on 16. I've looked through the logs but nothing seems to stand out. Looks like today has been a bad day...
A few things I have noticed when I suffer from one of these "Zigbee Storms".
They sometimes seem to be associated with the Zigbee Coordinator changing state (that is, when the Coordinator (hub) is re-booting, power down, or having some other issue. This may be related to what @danabw mentioned above. Hard to tell.
The reason why I suspect a wired device, is that when these "storms" happen, the relays in a bunch of my in-wall outlets click madly -- it sounds like the house is possessed. This is a strange thing, as none of these outlets have direct associations, so with the Coordinator down, where are the commands to turn outlets on and off coming from? My thought is that some of my hard wired devices must be freaking the heck out and spamming the network with random commands for a few seconds.
Why? Who the heck knows. It's certainly not the hub.
I've also noticed that sometimes (not often) a battery powered device seems to be a trigger -- say, when it's battery level gets low enough that it's ALMOST no able to transmit, I suspect that it's possible for such a device to send malformed commands or state, which the hub then translates into something else, which triggers the Zigbee outage.
Unfortunately, I don't have anything like a Zigbee Protocol Analyzer that I could park on the network and just record an outage. I think it might be interesting to see an outage like what @RobDaBush illustrates (this looks exactly like what I call a "Zigbee Storm") in a protocol analyzer...
Thanks for the replies. I had a bit of a meltdown last night and performed a soft reset to try to get to release .193. I ended up with the soft reset as I could not restore the database to .193. No matter what I did, which probably had a few mistakes in it, it would come up to build .200(my latest). So my backup was corrupt, I somehow renamed it to the wrong version, don't know. The reason for that is I noticed that .193 was up for a month and no reboots.
I updated as there was something in the update that I wanted/needed, don't remember.
Any who, channel 20 was my channel from the start. I came to that channel way back when and I had a C-4 and some issues with my Hue bridge. I've seen staff post that they've seen more problems on that channel but it always worked for me. I'm currently on 20, power 4. I've had 4 reboots since about 5 yesterday. I just bumped the power to 8 but I've been down this road, been there, done that. I've read from support staff that upping the power is actually counter intuitive as it can make it more susceptible to interference. Alot of my devices really like the channel as well. While channel hopping I had several instances where some of my devices would not work on some channels.
I don't. I mentioned I removed a device that I thought might be the issue and that was an Osram/Lightify lightstrip that I've had for about 9 years. That was my first suspect as I would have a wonderful looking zigbee map and then that thing would throw up all over it.
I removed it a while ago.
I have 6 repeaters. All but 1 are the old Iris Centrlite 3210-L. If they are the issue, it's not showing up in the graphs or logs.
The other is a SONOFF SNZB-06P mmwave sensor. There are a few people running them and I'm not aware of any issues. The driver just updated as I type this. Not that it means anything, I just thought it was funny timing. I just took a look a the graph and it has 2 red connection lines that I don't recall seeing before.
I just realized that something has gone wrong with my restore. It did not show it was on an older release but all the data is from an older database. It's showing devices that I renamed a while ago. Off to make things worse. Thanks.
Restoring to the database from .193 doesn't make the platform .193. You can use your existing database. Go to yourhubip:8081 and do theplatform rollback from there
Cranking up my zigbee power did not help, I have had a couple of resets today. I do have a lot of repeater devices. I'll try turning the zigbee power down. See how we go.
Thought I'd post to possibly help anyone chasing this issue. After many, many months of channel hopping, power settings and device testing. I think I finally have a stable system. I'm right back where I started and have been for many years on channel 20, power 8. I have come to the conclusion that Sonoff powered repeating devices do not like my setup. Many others are using them without issue but they do not like something in my setup or something in my setup does not like them.
The really annoying part was there was nothing that stood out in the logs with regards to these devices. They were a little chatty but nothing excessive. I removed my Sonoff ZBminil2 and ZBMicro(previously removed) devices from my system. I also have a Sonoff Presence sensor(snzb-06p) which does seem to cause an occasional zigbee reboot but they are infrequent, so I don't care as I really need this device. I thought about just getting another presence sensor but we'll see.
I was running them with @kkossev wonderful drivers but the issue persisted. I am running the presence with it right now. But that device somehow seems to go haywire on occasion and I just repair it and away we go. The last 2 times it happened this device managed to generate a ridiculous number of Hub Actions. I cannot find my other screen shot but it came in at over 1200 Hub Actions. I don't recall how long it was up.
On a side note, end devices are fine. With the exception of the Sonoff SNZB-02D Temp/ Humidity Sensors (with display). I cannot for the life of me get these paired. I have 3 Sonoff Temp/humidity sensors that paired without issue, just not the one with the display. I have 2 and can't get either paired. I have tried everything, pair avoiding repeaters, close to the hub, etc. They just won't pair. So, there it is. Good luck.
Edited for @danabw's whining, I mean viewing pleasure.
Another edit: I found the other screen shot. It was up for 2 days to hit that. I should also point out that this only happened recently after removing the other Sonoff repeating devices when trying to track down the latest infrequent reboots. I did not see anything like that during the months of trouble shooting.
