Vaibhav Singh

Even though I use GCP’s scheduled snapshot feature as a means to backup entire OS, for a last known good configuration for my compute instances. But it seems excessive in some cases because it doesn’t make sense to restore the entire OS for something as simple as recovering an accidentally deleted file.

Therefore, we must set a cron job which executes at regular intervals and archives the directory we care for most.

Create a shell script backup.sh in your home directory. Also create the target directory where backups will be stored.

#!/bin/bash
DATE=$(date +%Y-%m-%d-%H%M%S)
BACKUP_DIR="$HOME/backups"
SOURCE_DIR="/target/source-directory"
sudo tar -cvzpf $BACKUP_DIR/backup-$DATE.tar.gz $SOURCE_DIR

Write another bit of code to delete older archives so as not to exhaust the storage. The code below looks for files which have a modified date older than 10 from today and deletes them.

find $BACKUP_DIR -mtime +10 -type f -delete

Put all the above into a single script and give it execute permission with chmod +x backup.sh

Schedule Automated Execution

Set a cron job to execute the script at regular intervals. Install and enable cron if your environment doesn’t have it already.

sudo apt update
sudo apt install cron
sudo systemctl enable cron
crontab -e
  # Run backup.sh at 4:30am everyday
  30 4 * * * /absolute-path/backup.sh

Debugging

It’s important to test the execution in a temp directory with junk data first before attempting a first pass on your critical files. I typically set flags -xv at the shebang to see what’s going on with the script at execution stage. Eg. !/bin/bash -xv