Added support for storing transcoder metadata in S3

Signed-off-by: Fred Heinecke <fred.heinecke@yahoo.com>

transcoder/.env.example

@@ -4,8 +4,6 @@
 
 # where to store temporary transcoded files
 GOCODER_CACHE_ROOT="/cache"
-# where to store extracted data (subtitles/attachments/comptuted thumbnails for scrubbing)
-GOCODER_METADATA_ROOT="/metadata"
 # path prefix needed to reach the http endpoint
 GOCODER_PREFIX=""
 # base absolute path that contains video files (everything in this directory can be served)
@@ -40,3 +38,26 @@ POSTGRES_SSLMODE="disable"
 #  If this is not "disabled", the schema will be created (if it does not exists) and
 #  the search_path of the user will be ignored (only the schema specified will be used).
 POSTGRES_SCHEMA=gocoder
+
+# Storage backend
+# There are two currently supported backends: local filesystem and s3.
+# S3 must be used when running multiple instances of the service. The local filesystem is fine
+# for a single instance.
+
+# Local filesystem
+GOCODER_METADATA_ROOT="/metadata"
+
+# S3
+# Setting this configures the transcoder to use S3 as a backend.
+# S3_BUCKET_NAME=my-transcoder-bucket
+# All environment variables supported by the AWS SDK for Go (v2) are supported:
+# https://docs.aws.amazon.com/sdkref/latest/guide/settings-reference.html#EVarSettings
+# AWS_ACCESS_KEY_ID=abc123
+# AWS_SECRET_ACCESS_KEY=def456
+# AWS_ENDPOINT_URL_S3=https://s3.my-ceph-rgw-deployment.example
+# Unless you're running on an actual EC2 instance, you should set this to true.
+# This will disable the SDK from trying to use the EC2 metadata service to get credentials,
+# reducing startup time.
+# If you are actually using an IAM role profile for authentication, this should be set to false
+# or be left unset.
+AWS_EC2_METADATA_DISABLED="true"