Automating Restic Backups with failure notifcations

I did not write this article, this is a living “fork” including modifications that works for me.

Thank you and Credit to: https://roote.ca/about/tutorials/restic-backups/

Reliable Backups on a Linux System with Restic and systemd

Backups are a critical part of ensuring your data stays safe. Losing data is not something anyone wants to experience and reliable backups are how to ensure this never happens.

Backup infrastructure is critical but shouldn’t require large amounts of time to maintain. A program to create backups should work quietly in the background while ensuring data integrity and safety.

Restic is a great backup tool with powerful features like deduplication and encryption. It supports B2 cloud as a storage backend allowing offsite backups for an inexpensive price. Restic is a CLI tool and doesn’t include any scheduled tasks to run to create backups regularly. This tutorial configures systemd to run daily backups, prune the backup repository and run data integrity checks periodically.

Creating the config file

This file contains the authentication details as well as any settings to provide to the Restic command. Save the file with restrictive permissions only readable by the root user to prevent leaking the credentials.

# /etc/restic-backup.conf
BACKUP_PATHS="/etc/restic.includes"
EXCLUDE_PATHS="/etc/restic.excludes"
RETENTION_DAYS=7
RETENTION_WEEKS=8
RETENTION_MONTHS=12
RETENTION_YEARS=10
B2_ACCOUNT_ID="<B2_ACCOUNT_ID>"
B2_ACCOUNT_KEY="<B2_ACCOUNT_KEY>"
RESTIC_REPOSITORY="b2:<BUCKET_NAME>:<REPO_NAME>"
RESTIC_PASSWORD="<REPO_PASSWORD>"
RESTIC_CACHE_DIR=/srv/dev-disk/tmp/restic-cache

Configure the include and exclude files to specify which files to backup.

# /etc/restic.includes
/srv/dev-disk/syncthing
/src/dev-desk/media
# /etc/restic.excludes
**/node_modules/**

Configuring systemd

This backup configuration consists of three systemd service units, ‘restic-backup’, ‘restic-prune’, and ‘restic-check’.

These three units each run a specific job to maintain the Restic repository. There are separate timer unit files to run the backup tasks periodically. Once the unit files are in place, use the systemctl command to activate and start the timer services.

restic-backup

The restic-backup.service unit handles creating the backup snapshots.

# /etc/systemd/system/restic-backup.service
[Unit]
Description=Restic system backup
Before=restic-prune.service
Wants=restic-prune.service
JoinsNamespaceOf=restic-prune.service restic-check.service
OnFailure=ntfy@%n.service

[Service]
Type=oneshot
ExecStart=restic backup --verbose --tag auto-backup --iexclude-file $EXCLUDE_PATHS --files-from $BACKUP_PATHS --cache-dir $RESTIC_CACHE_DIR ; /usr/bin/sleep 20
EnvironmentFile=/etc/restic-backup.conf
SuccessExitStatus=3

The backup frequency is configured with the restic-backup.timer unit file.

# /etc/systemd/system/restic-backup.timer
[Unit]
Description=Backup with restic daily

[Timer]
OnCalendar=daily
RandomizedDelaySec=6hours
Persistent=true

[Install]
WantedBy=timers.target

Systemd timer’s will automatically associate with the corresponding service using the file name.

restic-prune

The restic-prune.service unit runs the prune command on the repo, using the retention periods configured in the restic-backup.conf file. This unit is configured to run after the backup unit.

# /etc/systemd/system/restic-prune.service
[Unit]
Description=Restic prune to clean up old backups
Requires=restic-backup.service
After=restic-backup.service
JoinsNamespaceOf=restic-backup.service restic-check.service
OnFailure=ntfy@%n.service

[Service]
Type=oneshot
ExecStart=restic forget --prune -o b2.connections=10 --compact --tag auto-backup --cleanup-cache --keep-daily $RETENTION_DAYS --keep-weekly $RETENTION_WEEKS --keep-monthly $RETENTION_MONTHS --keep-yearly $RETENTION_YEARS --cache-dir $RESTIC_CACHE_DIR
EnvironmentFile=/etc/restic-backup.conf

restic-check

The restic-check.service unit runs the consistency check on the repo to ensure data integrity. It’s configured to use the local cache in order to reduce the download API calls to B2 cloud to lower costs.

# /etc/systemd/system/restic-check.service
[Unit]
Description=Restic system backup repository consitency check
Conflicts=restic-backup.service restic-prune.service
JoinsNamespaceOf=restic-backup.service restic-prune.service
After=restic-prune.service
OnFailure=ntfy@%n.service

[Service]
Type=oneshot
ExecStart=restic --verbose=3 check --with-cache --cache-dir $RESTIC_CACHE_DIR
EnvironmentFile=/etc/restic-backup.conf

This unit is configured to run weekly.

# /etc/systemd/system/restic-check.timer
[Unit]
Description=Check restic repository consistency

[Timer]
OnCalendar=weekly
RandomizedDelaySec=1day
Persistent=true

[Install]
WantedBy=timers.target

Getting notified on failure

When a service fails, I want to know about it

Using the amazing, FOSS service of ntfy, we can create a systemd unit which POSTS to ntfy.sh/our-topic when a service fails.

# /etc/systemd/system/ntfy@.service
[Unit]
Description=Notification Service for %i to indicate failure

[Service]
Type=simple
ExecStart=/usr/bin/curl ntfy.sh/our-topic -d "%i"

Now add OnFailure=ntfy@%n.service to the [Unit] section of the units which should send failure notifications.

Receiving notifications

On LineageOS/Android you can subscribe to topics with the ntfy app.

Get it on F-Droid

Activating systemd units

After configuring the unit files the systemctl configuration will need to be reloaded, then the timer units activated and started.

systemctl daemon-reload
systemctl enable restic-backup.timer restic-check.timer ntfy@.service
systemctl start restic-backup.timer restic-check.timer

Optionally, run the service now to test it operates as expected

systemctl start restic-backup.service

The system should now create daily backups, prune any extra data daily, and check the repository consistency weekly.

Check the backup timer status with systemctl list-timers and individual units with systemctl status restic-backup.

Related Posts

© 2017 - 2022 · Home · Powered by Hugo ·