Belated, but intending to get this out next week, even with the work week. Please create train-22 branches for fxa-content-server and fxa-auth-server, when changes are in. I'll take care of fxa-auth-db-server and fxa-customs-server. Thanks.

https://github.com/mozilla/fxa-content-server/tree/train-22 7ff9e196311cef2e9e5a1726aa49081f2c0167a5

I've built rpms and new stacks in stage and switched DNS to them. The fxa-auth-server is a snapshot of current HEAD on master (ignore the 1.21 and look at the git SHA), since not sure if dcoates has anything to add. $ curl https://api-accounts.stage.mozaws.net/; echo ; curl https://accounts.stage.mozaws.net/ver.json; echo {"version":"1.21.0","commit":"0b8fbd8824a47cf729c5aad14a9b7b203200e6eb"} {"version":"0.22.0","commit":"7ff9e196311cef2e9e5a1726aa49081f2c0167a5","l10n":"8ec41f0a6f","tosPp":"2facea5477"} I've got one odd test failure that is only happening in stage, not on travis or locally that I'm looking into at the moment.

> I've got one odd test failure that is only happening in stage Oh, I see what this is, and it's not a real problem, but need to adjust a configuration in the tests.

https://github.com/mozilla/fxa-auth-server/tree/train-22 49aeb54cfbd5de503b15164fbbe62e3f7c365921 The only real difference is the basket.js process which can be enabled separately from the train.

Built fxa-auth-server rpm and updated stack in stage $ curl https://api-accounts.stage.mozaws.net/; echo ; curl https://accounts.stage.mozaws.net/ver.json; echo {"version":"1.22.0","commit":"49aeb54cfbd5de503b15164fbbe62e3f7c365921"} {"version":"0.22.0","commit":"7ff9e196311cef2e9e5a1726aa49081f2c0167a5","l10n":"8ec41f0a6f","tosPp":"2facea5477"}

Profile images seem borked: https://accounts.stage.mozaws.net/settings/avatar I suspect that it could be one or more of the following: 1) profile server needs a deploy too 2) oauth server may too 3) stage may need to be configured with clients with implicit grant permission: https://bugzilla.mozilla.org/show_bug.cgi?id=1064505 However, I believe that the OAuth server in stage is just down. I'm getting a 503 for all requests.

oauth and profile server experienced network errors about 2am and then couldn't reconnect to database. I restarted both and they are now back in service. WIll check state of avatar in a bit.

Getting "unknown client", which can probably be remedied by #3 in my above list.

:jrgm, we also need to add a new config param for this train, "trustedJKUs". Related issues: https://github.com/mozilla-services/puppet-config/issues/888 https://github.com/mozilla/fxa-dev/issues/58 https://bugzilla.mozilla.org/show_bug.cgi?id=1052883#c8 https://bugzilla.mozilla.org/show_bug.cgi?id=1052883#c9 has the values to add for dev and prod. It's not clear what to use for FxA stage.

I updated, in stage, the profile server and oauth server in stage to versions: $ curl https://oauth.stage.mozaws.net/ { "version": "0.20.0", "commit": "628597c904ba3515f695f1be4bfcf9bca8dbe2cc" } $ curl https://profile.stage.mozaws.net/ { "version": "0.22.0", "commit": "9f8fe75df765b8d9cb37211993018afa590ec0c8" } The oauth version is a snapshot of HEAD, and needs to get a tag for 0.22.0