So we had a 4 hour power/internet outage last evening. During the power/internet outage the st hub v2 was connected to a working UPS and i even put fresh batteries in it. the internet for the neighborhood was out so anything st did not work.
when the power and internet returned I noticed the led light changed form blue to green. I tried a couple buttons that control plugs, but none of that worked.
upon checking in the ide i see ghost id’s in the path for the zigbee buttons and plugs.
what i dont understand is why did these ghost id’s happen when the hub was powered during the power outage, but no internet.
How to delete a ghost --ONLY if this is an actual ghost, meaning how to clear a device ID out of your system that’s no longer associated with any Device - the process is basically the same for ZWave or Zigbee Just ZWave has 2 digit Hex numbers and Zigbee has 4 digit IDs
Devices CAN change IDs when rejoining. The difference between ZB and ZWave is Zigbee devices try to rejoin automagically and if you have ‘secure rejoin’ turned off they will. They’re SUPPOSED to inherit the old device when they do - and sometimes if things aren’t right they get a new ID. You won’t get more of an explanation than that and honestly there’s not much you can do about it except be able to recognize the issue and fix it.
so im running zwrepair and i see this “Network repair for Aeotec Repeater/Extender 1 [06]: Could not update neighbors” what does that mean and will it repair itself?
SmartThings has an unusual architecture because of its dependence on the cloud. For example, it is the cloud which marks a device offline.
Consequently when the hub and the cloud are not able to communicate with each other for a while, sometimes things can get weirdly out of sync. There have been multiple reports of this in the forums.
I know you don’t like to engage in theoreticals, so I’ll leave it at that.
so to do a zigbee heal i would unplug the hub (remove batteries) for 30 mins. this makes the zigbee endpoint devices panic and start looking for a new route?
Personally I wouldn’t put the battery back in after. It doesnt add much value and just increases the risk of a battery leak (used to be a big issue for v. 2 hubs)
They will start looking for a new route after you put the hub back on power once the 30 minutes is up. (And 20 minutes is probably enough.) Just make sure that all of the other Zigbee Devices are on power while the hub is off. Otherwise they won’t know the hub was missing and they won’t ask for new routes when it does come back.
The process of getting new routes can take time. It’s not instantaneous. When I was a field tech (for a different company, but working with Z wave and zigbee), we used to wait until the next day to check the new routes. They will probably show up sooner than that, but it’s a little unpredictable, especially if you have battery powered devices in your network, so it’s easier to just be patient.
Heh, I remember hearing that and never put the batteries in until yesterdays power outage in case the UPS went out. now it appears that w/o the internet the UPS/batteries didnt help.
when the motion sensor way out in the mailbox goes offline, its usually about 12-24 hours before it finds its new path.
so there is the zibee heal or repair way … i deleted a centralite outlet and now it wont add back in - the app times out. this is what is experienced in the past and just threw the towl in and scrapped everything to start over. this time I want to see if I can fix things - learn more so to speak.
Well, in truth once 35.x enhances local execution next week I can actually see a purpose… IF all the devices in the home are also on some kind of power system to survive the outage.
And, FWIW… I also have a V2 - Mine’s on an actual UPS. I wont put batteries in mine.
This (without the magic changing IDs) VERY eerily similar to what ive been fighting the last two weeks. Something (first itnwas an automatic WiFi channel change, then a few other things) stomped Zigbee and it took 12 hours to recover. It used to not take that long. (pre 32.x)
The more I worked on it the worse it got. I’m very suspicious something changed im the Zigbee stack in 32.x that is making zigbee routes take forever to recover.
I learned to be patient, wait hours, not minutes and the system would eventually recover. Not optimal. But it works great…
…except until it doesnt.
Unfortunately three months of solid performance is completely forgotten with one night of the wife not being able to turn off lights before bed.
By any chance do you have mains power Zigbee devices/repeaters that went offline during the power outage? If a device was routing through repeater that went missing and couldn’t reach the hub directly it would have an issue.
Also, I have found that when Zigbee devices rejoin with a new Device Network ID they will only reconnect to the previous device entry if the 16 digit Zigbee ID (e.g. 000D6F000575D915) is there. Otherwise ST has no way to reconnect it.
IMO you just need to be patient when these things happen, let the MESH heal and ST do its thing. I my 5 years with ST I have had power outages and had never had to start again.