Backups

Danger

No managed backup or disaster recovery systems exist for cluster storage. Users must independently back up critical data.

Setup Guide

1. AWS Configuration

  1. Create IAM user with programmatic access in AWS Console

  2. Generate S3 bucket in US-WEST-2 (Oregon region)

  3. Save your Access Key ID and Secret Access Key

2. GPG Encryption Setup

gpg2 --gen-key

Accept defaults, provide name, email, and passphrase.

Note the public key ID from output (e.g., pub 2048R/AE776EB5).

3. Install Backup Scripts

git clone https://github.com/zertrin/duplicity-backup.sh.git
cd duplicity-backup.sh
cp duplicity-backup.conf.example duplicity-backup.conf

4. Configure

Edit duplicity-backup.conf with:

  • ROOT - Directory to backup

  • DEST - S3 bucket path (e.g., s3://my-bucket/backups)

  • AWS credentials

  • GPG encryption keys and passphrase

  • Email notification settings

5. Create Scratch Directory

mkdir -p /resnick/scratch/$USER/duplicity-temp

6. Test Backup

./duplicity-backup.sh --backup

7. Automate with Cron

crontab -e

Add (runs daily at 3:41 AM):

41 3 * * * /path/to/duplicity-backup.sh --backup

Verification

Periodically test recovery:

./duplicity-backup.sh --restore /path/to/restore

Alternative: Rclone

See Rclone Backups to AWS for a simpler alternative.