My backup strategy 9 Jan 2014

For a long time I didn’t have a real backup strategy. I just had some HDDs with - hopefully - more than one copy of the important files. Then I decided to not trust into such a more or less random backup and build up a software RAID (RAID 5) using mdadm. Yeah, I know RAID isn’t a real backup but it’s a lot better than just having single HDDs which could fail and then the data is gone. After a while I wrote a little script to push essential files to the universities file system - the first solution that was allowed to name itself backup.

Update: I experienced that for longer running backups the script gets killed and found a solution here. I updated this post according to this.

Hardware

As my data grows it was time for the next step - a backup with more capacity. So I bought an external hard drive and a timer clock. Put the latter into plug-socket, setting up the timer for an suitable time slot, connect the power supply of the external hard drive to the timer clock and the USB into the PC - ready to set up the software part.

The actual backup is done using rsnapshot (Ubuntuusers wiki - german). To invoke this a litte helper script is called using a udev rule and a systemd service.

systemd service

The actual script has to be started as an systemd service. Otherwise it will be killed by systemd/udev. I created following service file /etc/systemd/system/backup@.service as described here.

[Unit]
Description=Backup to USB HDD
BindsTo=dev-%i.device

[Service]
Type=simple
ExecStart=path/to/Backup/backup.sh

udev rule

I used following udev rule (located in /etc/udev/rules.d/10-usb-hdd-backup.rules). The rule needs to be applied before the udisks2 rules (on ArchLinux the rule is named 80-udisks2.rules).

SUBSYSTEMS=="usb",ACTION=="add",KERNEL=="sd?1",ATTRS{serial}=="aabbccdd",
SYMLINK+="backup",RUN+="/usr/bin/systemctl --no-block start
backup@%k.service",	ENV{UDISKS_IGNORE}="1"

This calls the script path/to/backup.sh right after the recognition of the usb device. SYMLINK+="backup" creates the device /dev/backup for an easier device handling and ENV{UDISKS_IGNORE}="1" disables the auto mount of udisks2.

To retrieve the serial just plug the drive in and look up the serial with following command (where /dev/sdf is the external device).

$ udevadm info -a -p  $(udevadm info -q path -n /dev/sdf) | grep serial

backup.sh

This is the script I used to invoke rsnapshot. It mounts the partition, checks for the folders daily.0, weekly.0 and monthly.0 inside of the destination folder to call the daily part once a day, the weekly part once every 7 days and the monthly once every 30 days. Then it unmounts the partition and suspends the USB device using this script.

#!/bin/bash
#
#  script to use rsnapshot with a (not always mounted) external hard
#  drive (this script gets called by an udev rule)
#  see http://morrisjobke.de/2014/01/09/My-backup-strategy/
#
#  Copyright (C) 2014, Morris Jobke <morris.jobke@gmail.com>
#
#  This program is free software: you can redistribute it and/or modify
#  it under the terms of the GNU General Public License as published by
#  the Free Software Foundation, either version 3 of the License, or
#  (at your option) any later version.
#
#  This program is distributed in the hope that it will be useful,
#  but WITHOUT ANY WARRANTY; without even the implied warranty of
#  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
#  GNU General Public License for more details.
#
#  You should have received a copy of the GNU General Public License
#  along with this program.  If not, see <http://www.gnu.org/licenses/>.

DEV_NAME=/dev/backup
MNT_NAME=/mnt/backup
BACKUP_FOLDER=${MNT_NAME}/wigrid.snapshots
SUSPEND_USB_SCRIPT=path/to/suspend-usb-device.sh

# create directory if needed
if [ ! -d "$MNT_NAME" ]; then
	mkdir "$MNT_NAME"
fi

# mount
mount -t auto "$DEV_NAME" "$MNT_NAME"

logger "rsnapshot started"

if [ ! -d "${BACKUP_FOLDER}/daily.0" ] || [ $(find ${BACKUP_FOLDER}/daily.0
	-maxdepth 0 -type d -mtime +1 | grep daily) ]; then
	# if no daily backup is available or daily backup is older than 1 day
	# then run daily
	logger "rsnapshot daily"
	rsnapshot daily
fi

if [ ! -d "${BACKUP_FOLDER}/weekly.0" ] || [ $(find ${BACKUP_FOLDER}/weekly.0
	-maxdepth 0 -type d -mtime +7 | grep weekly) ]; then
	# if no weekly backup is available or weekly backup is older than 7 days
	# then run weekly
	logger "rsnapshot weekly"
	rsnapshot weekly
fi

if [ ! -d "${BACKUP_FOLDER}/monthly.0" ] || [ $(find ${BACKUP_FOLDER}/monthly.0
	-maxdepth 0 -type d -mtime +30 | grep monthly) ]; then
	# if no monthly backup is available or monthly backup is older than 30 days
	# then run monthly
	logger "rsnapshot monthly"
	rsnapshot monthly
fi

logger "rsnapshot finished"

#unmount
umount "$MNT_NAME"

# suspending USB device
# from https://raw2.github.com/vain/suspend-usb-device/master/suspend-usb-device
${SUSPEND_USB_SCRIPT} "$DEV_NAME"

I’m open for suggestions to improve this.

Bonus: You can check the current backup run with systemctl status backup@sdX1.service.