s3sync

Info

S3sync is a Docker container which backs up one or more folders to S3 using the aws cli tool. This is forked from joch's original S3Backup container but has been modified to make use of the aws cli rather than the outdated s3cmd. This has been also updated to Python 3.

Usage

This container makes use of the aws s3 sync command.

To tell s3sync what to back up, mount your desired volumes under the /data directory.

s3sync is configured by setting the following environment variables during the launch of the container.

Env variables

Cron schedule

Files are by default backed up once every hour. You can customize this behavior using an environment variable which uses the standard CRON notation. CRON_SCHEDULE is set to 0 * * * * by default, which means every hour. I would advise against setting this to a more frequent schedule.

Example invocation

Simple run

This will sync your /home/user local directory every hour to your specified S3 bucket, leaving everything else in the bucket intact.

docker run \
-v /home/user:/data/user:ro \
-e "ACCESS_KEY=AWS_ACCESS_KEY_HERE" \
-e "SECRET_KEY=AWS_SECRET_KEY_HERE" \
-e "REGION=eu-central-1" \
-e "S3PATH=s3://BUCKET_NAME_HERE" \
whatname/docker-s3sync

Advanced run

If you want more customization on the S3 side, you can use the S3SYNCPARAMS to input aws s3 sync CLI parameters such as --delete. You can also specify deeper paths in S3, and a cron schedule.

This will sync both the /home/user and /opt/files local folders to s3://your-bucket-name/this_prefix/ and delete everything else that's inside that prefix. Upon sync, the contents of s3://your-bucket-name/this_prefix/ will only be the two folders [and their files] that you just synced.

docker run \
-v /home/user:/data/user:ro \
-v /opt/files:/data/files:ro\
-e "ACCESS_KEY=AWS_ACCESS_KEY_HERE" \
-e "SECRET_KEY=AWS_SECRET_KEY_HERE" \
-e "REGION=eu-central-1" \
-e "S3PATH=s3://BUCKET_NAME_HERE/this_prefix" \
-e "S3SYNCPARAMS=--delete" \
-e "CRON_SCHEDULE=* 0 * * *" \
whatname/docker-s3sync

Unraid

s3sync is available on the Unraid Community Applications. See the Support Thread for more information, or search for s3sync in CA.

Future improvements (to-do)

Ability to assume role automatically

You can also use an IAM role that the user with the provided AWS Access Keys should assume to perform the sync. If not present, the pair of access and secret keys will be used directly.