After spending some quality time with jrgm and awsbox yesterday, I've reached the following conclusions:
* deploying 123done into stage can be done with a simple awsbox, since it doesn't need to share the stage.mozaws.net domain with the server stack -- it's just a relying party. * configuration of 123done-stage.lcip.org (tentative name) can be done with awsbox's -x command and a minor code change to 123done, or we can create a deployment branch with the config built-in. * however, deploying the oauth server into staging should be integrated with a process which can set up hosts in the stage.mozaws.net domain -- either CloudFormation or dcoates's mechanism. The name oauth.stage.mozaws.net makes sense here. * I have not investigated dcoates's mechanism closely. Danny, do you have any time in the next week or so to discuss it, and possibly walk me through a deploy? I will be filing the following bugs, mostly for tracking (and possible discussion) purposes: * On fxa-oauth-server, a 'question' bug to determine how to repeatably deploy an OAuth server into the stage env. (If this is better filed in Bugzilla, please feel free to speak up.) * On 123done, a bug to make the necessary changes to config management (or to the config itself) to deploy it against the stage env. Comments and discussion welcome -- is there anything large I'm missing? Thanks again, --KT. ----- Original Message ----- From: "John Morrison" <[email protected]> To: "Sean McArthur" <[email protected]>, "Karl Thiessen" <[email protected]> Cc: "Gene Wood" <[email protected]>, "Vladislav Filippov" <[email protected]>, "Daniel Coates" <[email protected]>, [email protected], "Benson Wong" <[email protected]>, "Chris Kolosiwsky" <[email protected]> Sent: Tuesday, June 3, 2014 7:46:43 PM Subject: Re: How do we intend to integrate the oauth-server and 123done into the FxA stage environment? (Resending, since autocomplete picked some other Ben... on my previous attempt) I know there has been been work on an rpm build script, a cloudformation config, and puppet-config rules, but haven't tried them myself. @mostlygeek or @ckolos would be able to comment more. https://github.com/mozilla-services/svcops/blob/master/services/firefox-accounts/fxa-oauth-server/mock-create.sh https://github.com/mozilla-services/svcops/blob/master/cloudformations/firefox-accounts/fxa-oauth-server.json https://github.com/mozilla-services/puppet-config/tree/master/fxa/modules/fxa_oauth On 06/03/14 19:06, Sean McArthur wrote: > I don't know much about our stage environment. I don't know how > involved deploying to stage is, or who does it. Without a better > understanding, I probably couldn't offer the best option. Thanks for > bring this up, however. > > > On Tue, Jun 3, 2014 at 4:51 PM, Karl Thiessen <[email protected] > <mailto:[email protected]>> wrote: > > > After a conversation with ckarlof, I thought I'd ask if anyone has > already thought about this before I start to formulate a plan: > > I think we're agreed that having an oauth server and a 123done > instance that talk to the staging env would be a good idea. My > question is, should we deploy them as an extension of what jrgm is > already doing on stage, or perhaps use Danny's whizzy new > stack-builder (which Vlad is happily hacking to accommodate the > two new servers)? There may be other alternatives I haven't > thought of, but those two seem to me to be the most likely candidates. > > Do we have proposed names? oauth.stage.mozaws.net > <http://oauth.stage.mozaws.net> and 123done.stage.mozaws.net > <http://123done.stage.mozaws.net> seem reasonable to me. > > Is this all already in the planning stages, so I'm fretting for > naught? > > Thanks for any context, > --KT. > > _______________________________________________ Dev-fxacct mailing list [email protected] https://mail.mozilla.org/listinfo/dev-fxacct
