Talk Nerdy To Me: Amazon EC2 backup pruning script

Update: here’s a text version without smartquotes. Download it, change the extension to .sh, and enjoy.

As a followup to this previous post - a script that automates EBS snapshots on an Amazon EC2 server - here’s a bash script to prune the EC2 snapshots created by the other script.

export EC2_PRIVATE_KEY={path to your AWS private key} 
export EC2_CERT={path to your AWS certificate} 
export JAVA_HOME=/usr/lib/jvm/jre
export EC2_HOME=/opt/aws/apitools/ec2

VOLUME_ID=/opt/aws/bin/ec2-describe-volume-status | cut -f 2 | grep -e "vol-[[:xdigit:]]\{8\}"
echo Volume ID is $VOLUME_ID

SNAPSHOT_ID=/opt/aws/bin/ec2-describe-snapshots --filter volume-id=$VOLUME_ID --filter description="Weekly backup" | sort -r -k 5 | sed 1,2d | cut -f 2
echo Snapshot ID to delete is $SNAPSHOT_ID

/opt/aws/bin/ec2-delete-snapshot $SNAPSHOT_ID
echo Snapshot deleted

This finds and deletes the oldest snapshot with a description “Weekly backup” (you’ll need to change that if your automated backup script generates a different description), while always keeping the two newest snapshots. If you want to keep more snapshots, change the sed command: changing it from sed 1,2d to sed 1,5d will keep the five most recent snapshots instead of the two most recent.

European elections? What European elections? Never heard of ‘em.