Not sure if this is just myself or not, every time HA reboots i’m having to re-authenticate Tuya (within 2 minutes of reboot) along with Google fit (which usually tells me it already has authorisation for what HA is requesting) and fitbit integration. perhaps a clash with something? any ideas?
1 Like
Same here; found your post trying to research it. Same two integrations as well; Tuya and Fitbit. Seems to have started for me on 2025.2.2 or .3 - not sure which.
Note that both reauths work every time, just not sure why they are failing on normal startup. In a previous issue, the tuya auth issues was said to be on their side and corrected, but to have the same integrations fail each time seems suspicious to me.
2 Likes
at least its not just me! annoyingn though, i think i was having similar issue few weeks back and was attributed to the HACS roborock integration previously, removing that and adding core version of roborock sorted it, just wondering if there is something else now causing it…
Well, I looked at the patch history to see if there were any changes that might cause this and then I looked at open issues and I don’t see anything on either source. I guess we might have to wait until someone else find this thread and reports similar issues. At that point I’d consider raising a GitHub issue but right now it looks like it’s just us. Which sucks.
At least it re-auths successfully each time.
i’ve noticed that some integration changes are not saving, have you noticed this? eg delete an integration and its still there on reboot, or delete a mobile app device and its back again after reboot
im getting a TypeError: Recursion limit reached in my logs, along with plenty other minor errors im trying to correct to reduce errors in my logs but not seeming to get anywhere
are you using the meross cloud integration at all?
I haven’t seen this but I haven’t had the cause to test it yet.
Yes, I am. Not sure if I see a connection though.
humour me, remove the cloud one and stick with the local version?
my issue is solved after making a few changes, and it seems to be the meross integration creating the problem as i havent had the tuya reconfigure since, along with some other issues that have now stopped. saw a thread mention similar with someone getting the recursion error
I’m not in a position to test this yet. Just got done with a ProxMox blowout; somehow I created a cluster unintentionally. Not sure how, I only ran a node repo update and because I was getting the subscription nag again, I ran the TTECK PM post install script, which I’ve run before w/o issue. Suddenly, I couldn’t manage a node without directly logging into it’s web gui. I could see the other node, but couldn’t do anything with it and started getting quorum warnings. I couldn’t use the shell/console and couldn’t start any of my containers or HA VM.
What a nightmare! After hours of struggling and research I managed to get the cluster config finally deleted and dead, but I can still only access a node directly. So it might be a few days before I can test this, but once I do I’ll get back to you.
So there’s an issue with the Meross integration and 2025.3.0 and as a result the integration wouldn’t load. But on rebooting after the core update I still got the authentication issues.
I went over to the Meross integration’s github to see if anyone else had the issue and if it was being worked on and found this post:
and it listed a work around which I deployed. Since there’s no reload option on the integration I had to restart HA to apply it. For the first time in days, I did NOT get the authentication error from either Fitbit or Tuya!
Now, it’s only been one reboot and I’ll have to try it again before I’m convinced it’s not just a coincidence, but you may want to check that post out and see if the same change works for you. Let’s keep in touch on this and see what comes of it.
Update: I just referenced this thread in the MerossIOT github issue I linked above. This way the devs there are at least aware of the possibility and can apply that to a permanent fix.
Another user who had similar issues to us:
also had their issues fixed with this work around! Looks like you were right; nice work man!
1 Like
Spoke too soon… I had to restart after applying the latest core update and it’s back again. I’ll try removing Meross entirely sometime this weekend and let you know if there’s any difference. Damn.