I recently started having a similar issue since around 0.35.X. Seems to happen randomly, but typically within a few hours of a fresh restart (under 24 hours) Wink stops updating the status. Same symptoms as described above, Light shows on when it isn’t / Wink App shows correct status. Door sensors stop responding, etc.
Only caught two pubnub error in the logs, going to keep an eye out for it again after the last restart.
16-12-27 14:14:03 homeassistant.components.wink: Error in pubnub JSON for Wink Relay hub polling API for current state
16-12-27 14:14:08 pubnub: take message interrupted: 'NoneType' object has no attribute 'loop'
I thought that it might have been related to using the Wink API configuration, as opposed to using access_token, but after I switched back to using access_token the same problem happens.
This is interesting. Keep an eye out for this again and see if it is the same error NoneType and the attribute loop that error is coming from the pubnub library so it may be something I’ll need to bring up with them. I’ll also check to see if there is a newer version of pubnub we should upgrade to.
Good catch, thanks.
Also that first line isn’t really an error, it just the best way I know how to address the way pubnub works. If they receive too many updates in a short amount of time they send back a JSON response with a URL to pull the state from and not just the state. So I just force a refresh from the wink api in this case.
I’m getting the same behaviour. Fresh restart works for some time then something happens to Wink. States are not properly reported. I can turn lights on and off but if they are on they always report as on. If off always report as off. Wink app shows proper status. I’ll take a better look at the logs but I don’t notice anything obvious so far.
16-12-28 07:53:15 homeassistant.components.wink: Error in pubnub JSON for Wink Relay hub polling API for current state
16-12-28 07:53:21 pubnub: take message interrupted: 'NoneType' object has no attribute 'loop'
I checked yesterday and there is a new version of pubnub, so maybe they have fixed this already?
I’ll update the pubnubsub-handler library to use the newest version and we canbsee if that fixes the issue? Ill look into opening a pubnub issue as well.
Strange I am not seeing this. Those of you that are, what devices are you using?
GoControl Z-Wave Home Security Suite (just the door / window sensors)
GE Z-Wave Wireless Lighting Control Outdoor Module
iHome WiFI Smart Plug
Remotec Zwave Dry Contact Fixture Module
I have a few other devices in Wink that I don’t really use in any capacity with HA via the Wink platform, but they are setup in Wink and get loaded by the HA component. I just hide them and don’t automate around them. Using Native platforms instead.
I have mostly Caseta dimmers and pico remotes. A few connected bulbs, 2 GoControl window / door binary sensors, 1 GoControl PIR, 1 Schlage connect lock, 1 z-wave leviton switch & 1 z-wave leviton plug.
It seems that GoControl PIR is consistently involved. For instance right now I just restarted HASS and all was working fine. Tripped PIR and now it shows as on permanently. I’ve disconnected it and will see if that helps. Wink not reporting properly either. In one spot it shows motion detected in another on same screen says motion stopped.
I just ran into the issue again, this time I was watching for it and it coincided with a network blip on my LAN. So I think you might be on the right track about pubnub failing to connect then not being able to reconnect.
16-12-28 12:20:54 pubnub: take message interrupted: 'NoneType' object has no attribute 'loop'
16-12-28 12:21:00 homeassistant.core: Error doing job: Task exception was never retrieved
I removed the GoControl PIR. Same issue happening. I must have been imagining the connection. I am getting this error too:
16-12-28 13:30:51 pywemo.ssdp: Error fetching description at https://192.168.1.106:8888
16-12-28 13:31:32 netdisco.ssdp: Error fetching description at https://192.168.1.106:8888
16-12-28 13:35:50 homeassistant.components.wink: Error in pubnub JSON for Office Remote hub polling API for current state
16-12-28 13:35:51 pubnub: take message interrupted: 'NoneType' object has no attribute 'loop'
106 is my Wink hub. Not sure if that is meaningful. I think I’ve been getting the netdisco error for a long time. The pywemo is new though.
The net discovery and wemo is because those components scan the network they get an invalid response from the hub (because it isn’t a wemo). Nothing to worry about.
Still really strange so many of you are seeing this, and I am not. Guess my network is solid lol.
I run mine on a VM connected via a virtual interface and switch on top of dual physical interfaces that are bonded. Part of a VM Farm. Occasionally I’ll see a warning on one of the physical interfaces (usually when I’m starting to saturate the network connections), which in this case I saw the error. The VM and Operating System don’t seem to ever notice having an issue.
Anyone else running a router they can’t reach? I can’t access the GUI for my netgear router (Running DD-WRT). I chalked this up to too many reboots a few days ago. So I took DD-WRT out of config and figured I would be able to access my router after it reset # of attempts but I’m still blocked. Very curious. Not sure if it’s related.
Got a response from pubnub, and they said there were some connection manager updates in the latest version and that we should give it a try. I will get this added to the library today or tomorrow and maybe someone that is having this issue can give it a try?
Also for those of you having the NoneType issue, what region of the world/US are you in? Pubnub was asking wondering if it is some DNS related issue somehow.
6:35 would have been about the time of the first state in Wink change this morning.
16-12-28 21:46:26 pywemo.ssdp: Error fetching description at https://192.168.1.106:8888
16-12-28 21:47:23 homeassistant.helpers.state: reproduce_state: Unable to find entity switch.christmas_tree
16-12-28 21:47:35 netdisco.ssdp: Error fetching description at https://192.168.1.106:8888
16-12-29 06:35:29 homeassistant.components.wink: Error in pubnub JSON for Remote hub polling API for current state
16-12-29 06:35:30 pubnub: take message interrupted: 'NoneType' object has no attribute 'loop'
Seems like all of us are using the GoControl PIR, could that be significant? Should also note that I restarted HA around 2:00 PM yesterday for unrelated reasons and I don’t see anything in the log so far.
Sorry to hijack here…so far I’m not seeing the issues in HA that you guys are, but I am seeing some weird behavior with my GoControl PIR seeming to cause other devices to go offline or report false alarms. For example my Ecolink PIR will report motion detected randomly, and doesn’t show stopped until someone actually walks in front of it. At the exact same time, my GoControl PIR in the basement always reports a temperature of 57 degrees (which is way below its normal readings). I’m wondering if any of you guys are seeing anything similar? Perhaps it’s all related??
