Completely reset Loaner data/Downgrading issues about loaner HOT 3 CLOSED

Thanks for reporting this. I can't provide any specific guidance for reverting or downgrading (usually we'd recommend against that since you're potentially affecting the Datastore entity model).

The current version we recommend is our master version. There was a push a while ago that caused some associated_fleet issues, but the latest master push shouldn't have this issue. Additionally when you're updating the serving version are you updating the serving version for each of the services (action-system, default, chrome, endpoints)?

Here is the 0.7.6a commit: 16df9a5#diff-eecff035f2a636bf3d7ef27d96f9d648

You should also flush memcache and indexing if possible to see if that remediates some of the issues.

If you are comfortable blowing away this whole deployment and starting from scratch (this is definitely the nuclear route), you'd need to delete everything in Datastore, flush the memcache, and clear the indexing. Then redeploy and setup as you normally would. Do keep in mind that if you take this approach, it will delete everything from the app (including devices).

from loaner.

Flushing memcached after the Datastore wipe allowed the bootstrapping to happen correctly and eliminated the errors that were appearing in StackDriver. However, I am now running into an OAuth issue in the Chrome app. The Chrome app appears to be unable to get an OAuth token. I triple-checked that the Chrome client ID is authorized correctly. I've attached a screenshot of the debug screen from the Chrome app.

from loaner.

The OAuth issue was resolved and was due to a misconfiguration with the client ID. Thanks!

from loaner.