Unable to stay logged in using new OAuth mechanism

[ASwingler]

Hi. Thanks for the new version v0.5.2. I'm struggling to stay logged in now. The integration appears to log me in correctly (I receive "Reconfiguration successful") but none of the sensors return data and I get a "Failed to set up: Missing refresh token or DPoP key" in the integration.

Here are the logs:

Logger: custom_components.generac.auth
Source: custom_components/generac/auth.py:746
Integration: Generac MobileLink (documentation, issues)
First occurred: 8:59:58 AM (3 occurrences)
Last logged: 9:02:34 AM

Generac auth: step=resume -> loc=com.generac.mobilelink.auth0://auth.ecobee.com/ios/com.generac.mobilelink/callback?code=***DELETED***&state=***DELETED***
Generac auth: step=resume -> loc=com.generac.mobilelink.auth0://auth.ecobee.com/ios/com.generac.mobilelink/callback?code=***DELETED***&state=***DELETED***
Generac auth: step=resume -> loc=com.generac.mobilelink.auth0://auth.ecobee.com/ios/com.generac.mobilelink/callback?code=***DELETED***&state=***DELETED***




Logger: custom_components.generac.auth
Source: custom_components/generac/auth.py:398
Integration: Generac MobileLink (documentation, issues)
First occurred: 8:59:58 AM (3 occurrences)
Last logged: 9:02:34 AM

Generac auth: step=password -> loc=/authorize/resume?state=***DELETED***
Generac auth: step=password -> loc=/authorize/resume?state=***DELETED***
Generac auth: step=password -> loc=/authorize/resume?state=***DELETED***




Logger: custom_components.generac.auth
Source: custom_components/generac/auth.py:379
Integration: Generac MobileLink (documentation, issues)
First occurred: 8:52:40 AM (7 occurrences)
Last logged: 9:02:33 AM

Generac auth: step=identifier -> loc=/u/login/password?state=***DELETED***
Generac auth: step=identifier -> loc=/u/login/password?state=***DELETED***
Generac auth: step=identifier -> loc=/u/login/password?state=***DELETED***
Generac auth: step=identifier -> loc=/u/login/password?state=***DELETED***
Generac auth: step=identifier -> loc=/u/login/password?state=***DELETED***




Logger: custom_components.generac.auth
Source: custom_components/generac/auth.py:266
Integration: Generac MobileLink (documentation, issues)
First occurred: 8:52:40 AM (7 occurrences)
Last logged: 9:02:33 AM

Generac auth: step=authorize -> 302 loc=/u/login/identifier?state=***DELETED*** set-cookie-count=4 jar-after=['auth0', 'auth0_compat', 'did', 'did_compat']
Generac auth: step=authorize -> 302 loc=/u/login/identifier?state=***DELETED*** set-cookie-count=4 jar-after=['auth0', 'auth0_compat', 'did', 'did_compat']
Generac auth: step=authorize -> 302 loc=/u/login/identifier?state=***DELETED*** set-cookie-count=4 jar-after=['auth0', 'auth0_compat', 'did', 'did_compat']
Generac auth: step=authorize -> 302 loc=/u/login/identifier?state=***DELETED*** set-cookie-count=4 jar-after=['auth0', 'auth0_compat', 'did', 'did_compat']
Generac auth: step=authorize -> 302 loc=/u/login/identifier?state=***DELETED*** set-cookie-count=4 jar-after=['auth0', 'auth0_compat', 'did', 'did_compat']

[binarydev]

Hey @ASwingler thanks for reporting the issue! I can't seem to recreate the problem, but I'll keep trying different scenarios to do so. Does it persist if you remove and re-add the integration entry?

[cheesedogf150]

Hey @ASwingler thanks for reporting the issue! I can't seem to recreate the problem, but I'll keep trying different scenarios to do so. Does it persist if you remove and re-add the integration entry?

I was getting the same issue as the OP. This fix worked for me.

[plynkus]

The same symptom occurred for me after first update/reconfiguration but a manual restart of HA afterward resolved the issue.

(Edit: Note that this was in addition to the automatic restart performed as a part of the integration update itself.)

[binarydev]

Thanks @plynkus that's helpful! @ASwingler or @cheesedogf150 let me know if an additional HA restart fixes this for you as well.

[ASwingler]

@binarydev The integration works correctly after a full reboot of HA. Leaving this case open in case you want a reminder to adjust the docs accordingly. Feel free to close.

[mattmeiser]

Second reboot did not work for me but removing and adding the integration back worked.

[binarydev]

Hey @sslivins, I saw you already got a head start on a fix. That's great and much appreciated! If you open a PR when ready, I'll test and accept before cutting a v0.5.3 release

[sslivins]

Thanks @binarydev! It's ready. I opened #286 (all checks green). Root cause was that async_setup_entry raises ConfigEntryAuthFailed on the v1->v2 upgrade before the update listener is registered, so the reauth/reconfigure credential write had nothing to trigger a reload. That's why everyone needed a full restart or a remove+re-add. The PR schedules an explicit reload when the entry isn't loaded, plus regression tests. Looking forward to v0.5.3!

[jr-ratliff]

THANKS FOR DOING THIS!!!!!

Working after 2nd restart of HA.
Thank you.

[binarydev]

Huge thanks again to @sslivins for the PR with the fix! v0.5.3 is now live, so this issue should be fixed for other users moving forward!