I’ve been having same issues for past couple days with that error in the log and Wink status not updating. At this point I think I’m finished jumping through hoops with Wink and am going to just change over to HA straight up. They are really blowing it, I was a big promoter…
Can confirm - I’m seeing fewer of these subscripe loop “connection aborted” errors in b4 though (so far).
Curious that @rpitera is not seeing the same lag in status updates. It’s really bad for me - seriously, 10 seconds before a UI switch returns to its normal/correct state.
Ok, that was what I thought, but didn’t know if there was a different log to look at. Still not pulling in the custom component, and still pointing to pubnub instead of notifier.wink.
Hmmm what are you running hass.io? Are you sure the layout is as follows? HA_CONFIG_DIR/custom_components/wink/* ?
Gary - What model hub are you on and are you Wifi or Enet?
I’m on a Hub2 with a direct Enet connection to my router. Might account for difference? Also running my hass.io in a VM on a multicore desktop.
Wink Hub 2 - Ethernet - plugged directly into Router, as is the HASSIO/RPI 3B+. So LAN network-wise, it is as optimal as can be I believe.
I’m researching various NUC’s as well as Zwave sticks or hub alternatives too.
EDIT: Now seeing the same amount of subscribe loop errors - about once every minute.
I concur, about the same frequency - every 60-80 seconds.
That seems to be what I see in the proxy with their official app as well so I guess it’s normal?
I suppose we could suppress it by setting the component error level to Critical or something, right?
UPDATE: Just checked my config and I am already at critical for homeassistant.components.wink…
Even my relays work now… I use them with nodered for smart lights !!!
Your’re awesome will !!!
I’ve also enabled Wink local control, but per Will’s post 85:
"4. Updates are still pushed via PubNub so internet is required for reliable updates."
It’s PubNub that’s throwing the error not wink.
Try setting pubnub to critical
Just to make sure I have the config right:
homeassistant.components.pubnub: critical
Is that correct?
Confirming this suppresses the pubnub error from the log. Note that the information page still shows a summary of the error (as well as “occurs xxxx times”), which is good. The log isn’t spammed.
logger:
default: warning
logs:
urllib3.connectionpool: error
pubnub: critical
Note that the connectionpool entry was needed to suppress the warning that was thrown from the original wink component. Not really needed with the custom wink component.
2 Likes
Thanks Gary - that fixes it for me too. Saved me some typing as well!
Coming in late… @w1ll1am23 I updated your custom that you sent earlier to this latest and changed the b3 to b4… no sensor check ins as yet… nothing being updated.
but I am seeing this in the logs…
Connection pool is full, discarding connection: notifier.wink.com
I have it working now! I needed to create a custom_components folder under config, then put the wink folder under there.
Is this because I customized an entity previously, by changing the icon? That is the only customization I have done so far in the customize.yaml file. Still trying to figure this all out.
customize.yaml has nothing to do with implementing custom components. That is used only for customizing the UI (icons, what is hidden, etc.).
You are on the right track now with wink in the custom_components folder for testing these ad-hoc fixes. This is temporary! Keep an eye out on the community thread for further developments, and if/when these updates stabilize, then a PR should be issued to make it “official”.
When you see the Wink component is mentioned in the HA release notes, then you can remove wink from the custom_components and update Hass the normal way.
1 Like
Not sure what that’s about. Is it still happening? Anyone else seeing the same thing as @jlvandusen?