Closed
Bug 1252141
Opened 9 years ago
Closed 9 years ago
Deploy TestPilot to Stage/Prod
Categories
(Cloud Services :: Operations: Deployment Requests - DEPRECATED, task)
Cloud Services
Operations: Deployment Requests - DEPRECATED
Tracking
(Not tracked)
RESOLVED
FIXED
People
(Reporter: mostlygeek, Unassigned)
References
Details
This bug will track progress of creating the ops assets (jenkins, cloudformation, puppet, etc) and deployment of the stage/prod environments for TestPilot.
| Reporter | ||
Comment 1•9 years ago
|
||
Stage has been created / deployed.
- stage deploys successfully
- jenkins, cloudformation and puppet code has been created
- secrets stored encrypted at rest with SOPs
It is available at https://testpilot.stage.mozaws.net.
Still need to verify that all configurations are set correctly.
| Reporter | ||
Comment 2•9 years ago
|
||
Testpilot fixed in stage.
- uses production FxA (accounts.firefox.com) for OAuth
- fixed configuration issues
I was able to sign in. \o/
| Reporter | ||
Updated•9 years ago
|
Summary: Deploy TestPilot to Stage/Dev → Deploy TestPilot to Stage/Prod
| Reporter | ||
Comment 3•9 years ago
|
||
Production deployed. Was able to log in.
| Reporter | ||
Updated•9 years ago
|
Status: NEW → RESOLVED
Closed: 9 years ago
Resolution: --- → FIXED
| Reporter | ||
Comment 4•9 years ago
|
||
Les, could you test out production and mark this bug as verified if it is ok.
Flags: needinfo?(lorchard)
Comment 5•9 years ago
|
||
Okay, my FxA login works, and the admin password works.
But, I get Secure Connection Failed errors on these admin pages when I try to create a new experiment:
https://testpilot.stage.mozaws.net/admin/experiments/experiment/add/
https://testpilot.firefox.com/admin/experiments/experiment/add/
Both of those pages involve multipart/form-data file uploads for images needed in experiment content. Other admin pages with form posts work, so I suspect the image submissions are the problem
Flags: needinfo?(lorchard)
| Reporter | ||
Comment 6•9 years ago
|
||
Looks like it was blowing up the max body size that nginx allowed. I've increased this to 256MB.
Stage and prod are both updated.
Comment 7•9 years ago
|
||
Okay, now I can get past the form submission on stage. But, the images 404 on S3:
http://testpilot-stage.s3-website-us-east-1.amazonaws.com/experiments_experiment/2/6/26faa091bc6f0767d90031ebb3898d6c_thumbnail_1457394529_0974.gif
Code: NoSuchWebsiteConfiguration
Message: The specified bucket does not have a website configuration
BucketName: testpilot-stage
| Reporter | ||
Comment 8•9 years ago
|
||
Did you want these S3 buckets to also serve HTTP?
That is no problem, in prod we prefer to put a CDN in front so the URL will have to be https://testpilot.cdn.mozilla.net
Comment 9•9 years ago
|
||
Though, FWIW, it does look like the image made it to S3:
http://s3.amazonaws.com/testpilot-stage/experiments_experiment/2/6/26faa091bc6f0767d90031ebb3898d6c_thumbnail_1457394529_0974.gif
Comment 10•9 years ago
|
||
> Did you want these S3 buckets to also serve HTTP?
Yes, these are image uploads for Experiment site content
| Reporter | ||
Comment 11•9 years ago
|
||
Alright, buckets have webhosting enabled.
Comment 12•9 years ago
|
||
Thanks Ben and Les. This is looking great.
Comment 13•9 years ago
|
||
Images on staging are working, but the URLs for prod images respond with 400 Bad Request:
http://testpilot-prod.s3-website-us-east-1.amazonaws.com/experiments_experiment/2/6/26faa091bc6f0767d90031ebb3898d6c_thumbnail_1457455911_0008.JPG
Code: IncorrectEndpoint
Message: The specified bucket exists in another region. Please direct requests to the specified endpoint.
Endpoint: testpilot-prod.s3-website-us-west-2.amazonaws.com
I can confirm that the image made it up to S3, though, with a direct bucket URL:
http://testpilot-prod.s3.amazonaws.com/experiments_experiment/2/6/26faa091bc6f0767d90031ebb3898d6c_thumbnail_1457455911_0008.JPG
Comment 14•9 years ago
|
||
Oh yeah, and this URL works:
http://testpilot-prod.s3-website-us-west-2.amazonaws.com/experiments_experiment/2/6/26faa091bc6f0767d90031ebb3898d6c_thumbnail_1457455911_0008.JPG
So, maybe just a matter of changing the configured base URL? Though, we probably want an HTTPS CDN in front of that so we don't have mixed content issues on prod
| Reporter | ||
Comment 15•9 years ago
|
||
You need to log in
before you can comment on or make changes to this bug.
Description
•