s3proxy

Simple HTTP server which proxies requests to Amazon Simple Storage Service. It allows legacy HTTP fetchers to download privately shared files from Amazon S3 buckets.

Usage

Use the following URL format to download a file from S3 bucket over s3proxy:

$ curl http://localhost:8080/s3-region-name/s3-bucket-name/path/to/your/file.txt

Deployment

Specify AWS credentials or EC2 Role Provider URL as environment variables and specify a port number as the only parameter:

$ AWS_ACCESS_KEY_ID=XXXX AWS_SECRET_ACCESS_KEY=XXXX s3proxy 8080

or

$ s3proxy 8080

which would have the same effect as

$ AWS_METADATA_URL=http://169.254.169.254:80/latest s3proxy 8080

Deployment on DC/OS

1. Edit s3proxy.marathon.json:
2. Specify credentials with env.AWS_ACCESS_KEY_ID and env.AWS_SECRET_ACCESS_KEY or EC2 Role Provider URL with env.AWS_METADATA_URL or do not specify env at all which would have the same effect as if you set env.AWS_METADATA_URL to http://169.254.169.254:80/latest.
3. Specify path to some small file which resides on S3 bucket with healthChecks.path in order to set health-check or remove healthChecks field if you don't want to have health checks.
4. If you'd like to use your own build of s3proxy, then change s3proxy executable URI with uris[0]. Default URI points to the file built for 64-bit Linux.
5. Deploy s3proxy using Marathon UI or DC/OS CLI:

$ dcos marathon app add s3proxy.marathon.json

Build

1. Install Go: https://golang.org/doc/install
2. Run

$ GOOS=linux GOARCH=amd64 go get github.com/adyatlov/s3proxy

Omit setting GOOS and GOARCH if you'd like to build s3proxy for the host platform.