-
-
Notifications
You must be signed in to change notification settings - Fork 649
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Use S3 Sync for CI deploys to avoid recopying files #7895
Conversation
We only want to use these values when using the script. Otherwise, AWS picks them up and tries to use them.
Going to run all CI via the branch build, not the PR build, since this impacts only non-PRs. I ran yesterday one shard on the PR build to confirm that this won't break PRs uploading |
Big thanks to @benjyw for all the help to land this! |
#7895 works for branch builds, but not pull request builds due to the error `Partial credentials found in env, missing: AWS_SECRET_ACCESS_KEY`. This would happen because we always would define `AWS_SECRET_ACCESS_KEY`, but the value `AWS_SECRET_ACCESS_KEY` would only be defined on branch builds because it's encrypted. Instead, we should only define `AWS_SECRET_ACCESS_KEY` if we're on a branch build.
Thanks! Relatively little should already be in S3, because the path is 1) unique to the SHA, 2) unique to the wheel that is being created. I think that only universal wheels could potentially be uploaded multiple times. |
We currently copy all of
dist/deploy
into S3, even if the file is an exact copy of something already in S3. This poses a real cost, both monetary and time.Instead, here we use
aws s3 sync
to only transfer files that are different from what is currently in the bucket.Will close #7258.