Search open search form


If you have valuable data that requires backup you may want to use the currently installed Duplicity backup client or another client/process of your choosing. Duplicity supports encrypted backups to many cloud storage vendors including Google, Amazon S3, Backblaze B2, Azure etc.
As an example, to configure a simple backup of a specific directory to AWS S3 perform the following. 

Create IAM user and associated AWS API keys in AWS Console to allow access to S3. (For simplicity sake this example gives full access to S3 in AWS. You may want to edit to conform to your specific AWS security requirements)

Log into your AWS console.

Click on 'Services' followed by 'IAM'. 

In IAM console click 'Users' and 'Add User'.

Username: duplicity-backups
Access: Programmatic access

Under Set Permissions click 'Attach existing policies directly', select 'AmazonS3FullAccess'

Click Next, Next (Review) followed by 'Create User'

Note the Access Key ID and Secret Access Key, you will use these in the configuration of Duplicity.

Create a S3 bucket in US-WEST-2 (Oregon) 

Create a GNU privacy guard (GPG) key to be used to encrypt the backups. 

Run 'gpg2 --gen-key' , select default key (hit enter), select default keysize (hit enter) , select default for key expiration (hit enter) then enter real name, email address and passphrase. 

You'll see the public key listed (line that starts with 'pub') similar to below. The key part (AE776EB5 in my case) will be used in the configuration of Duplicity. 

pub   2048R/AE776EB5 2019-04-08

Pull down and edit helper scripts to ease configuration of Duplicity

cd ~/
git clone
cp duplicity-backup.conf.example  duplicity-backup.conf

At minimum, edit the following entries in the duplicity-backup.conf file. Do not include the notes in parentheses.

ROOT='/home/<your-username>/data'   (directory to backup)
HOSTNAME=$(hostname -f)   (preexisting-s3-bucket)
AWS_ACCESS_KEY_ID="AKIAX4EICHL6ZQXSXXXX"    (IAM access keys for bucket)
STORAGECLASS="--s3-use-ia"    (optional: uncommenting this will use S3's infrequent access storage class to save on cost)
PASSPHRASE='<your GPG key passphrase selected during GPG key creation>'
GPG_ENC_KEY="B20EE8XX"    (public GPG key ID)
GPG_SIGN_KEY="B20EE8XX"   (public GPG key ID)
STATIC_OPTIONS="--full-if-older-than 14D"
LOG_FILE="duplicity-$(date +%Y-%m-%d_%H-%M).txt"
EMAIL_TO=""    (replace with your email address)
EMAIL_FAILURE_ONLY="no" # send e-mail only if there was an error while creating backup
MAIL="mailx"     # default command for Linux mail
#ECHO=$(which echo)Save the conf file.    (This is only used to debug)

 If the TEMP_DIR above does not exist you must create it before the first backup is run. See documentation links at the end of this page.

mkdir /central/scratch/<your-user-name>

Install Duplicity

(Central cluster users should skip this step)

The script above requires the duplicity client. You can download the client from 

Install the S3CMD client.

(Central cluster users should skip this installation step) 

Duplicity requires S3CMD in order to show remote file sizes. 

yum install s3cmd

Configure the s3cmd client. 

(Central cluster users will need to perform this step)

s3cmd --configure

enter the same AWS access keys used in the duplicity.conf

accept defaults for everything else (hit enter) 

Run a test Backup.

Now run a test backup to assure everything works as intended. 

p.p1 {margin: 0.0px 0.0px 0.0px 0.0px; font: 13.0px Menlo; color: #000000; background-color: #fdfdfd} span.s1 {font-variant-ligatures: no-common-ligatures}

/home/<your-username>/ -c /home/<your-username>/ --s3-use-multiprocessing -b

Create a crontab entry to automate backup. 

crontab -e

Add the following link to run the backup once per day at 3AM. You should receive a email message from cron whenever the backup script is run. Be sure to periodically check the script to verify if your backups are running correctly and run test recovery often to assure you can properly recover and decrypt data from the backups.

41 3 * * * /home/<your-username>/ -c /home/<your-username>/ --s3-use-multiprocessing -b

Save the cron file. This will run your backup every day at 3:41AM.