A data backup is a copy of computer data taken and stored somewhere else so that it may be used to restore the original after a data loss event.
Backups can be used to recover data after its loss on the original computer due to deletion or corruption, or backups can be used to recover data from an earlier time. The last case should be avoided and replaced with revision control systems.
There are different kind of backups depending the data to be backed up. A backup Of system data can be used to recover a system of well planned, called disaster recovery. While a backup of user data can recover errors (deletion for example) made by users.
No all backup systems can be uses as a disaster recovery. And sometimes the ability depends on the complexity of the system to backed up: Computer clusters and database server for example.
A backup system should only contain data considered worth saving and at least one copy of it. Some systems are able to provide more than one copy or can reconstruct data depending on a time stamp. They are suited for user data recovery.
Data storage requirements of backup systems can be very large. Due this some solutions overcome the price tag by offering expensive primary storage and less expensive secondary storage which often differs in retrieval speed. Examples are: SSD for primary storage and tapes for secondary storage.
Data to be stored needs to be selected. The selection can be up on time or location or other criteria. The backup systems has to support certain methods for different data types as: databases, open files, compression, encryption and de-duplication.
Every backup should be planned with a strategy in mind. The 3-2-1 rule states, there should be 3 copies of the data, stored on 2 different media types (of 2 different computers) and 1 copy should not be off site.
The reason why one should use at least 2 different computers is, that sometimes even hardware (not the media) might introduce errors to the backup. A defective raid controller might not only crash the file systems of spinning hard disks attach to it but also that of SSDs, for example.
Usually the smallest entity to be backed up is a file. Files are organized via a file system. The data types are: static data (files, …), live data (database, log file, …) or meta data (boot sector, partition data, file modification data, …).
For example, to back up the live data of a database it is often required to shut down the database to make the database file static, before making a backup.
Every Linux is shipped with backup tools.
They can be used together with crond and/or ssh for example to create a small and reliable backup system.
For example, if you have mounted a directory of an external server via NFS to /remote/server/dir you can put this script under /etc/cron.daily to back up the /etc directory.
#!/usr/bin/bash
DIR=/remote/server/dir
FILE=$DIR/backup-etc-`date +'%F'`.tar.gz
if [ -d $DIR ]; then
cd /
if [ -f $FILE ]; then
echo "ERROR: File $FILE exists"
exit 2
else
/bin/tar czf $FILE etc
fi
else
echo "ERROR: No directory $DIR"
exit 3
fiThis can of course not be used to make a disaster recovery of the system, but parts of the system configuration could be restored.
If the target is to backup a partition (which might be questionable), so called snapshots can be used for certain file systems. The following list should be considered with caution as it do not claim that the mentioned file systems are stable. Sometimes the can be done via file system specific fsck tools, external commands or via an interim layer, like LVM.
Software that can handle LVM:
Software that sounds like a snapshot, but isn’t:
rsync and hard links. Debian-Edu (Skolelinux) use rsnapshot for backups in school environments.While it is advisable to create once own backup systems for young system administrators to understand the underlying problems of backups, there are free open source software already provided to Linux that can be used, if one understand the core principles of backups.
| Package | Debian | Interface | Debian Buster | Bullseye | Bookworm |
|---|---|---|---|---|---|
| amanda- | CLI | 1:3.3.9-5 | 1:3.5.1-7 | 1:3.5.1-11+deb12u1 | |
| Areca Backup | n.a. | ||||
| BackupPC | GUI | 3.3.1-4 | 4.4.0-3 | 4.4.0-8 | |
| BackInTime | GUI | 1.1.12-2 | 1.2.1-3 | 1.3.3-4 | |
| Bacula | 7.4.4+dfsg-6+deb9u2 | 9.6.7-3 | 9.6.7-7 | ||
| Bareos | Buster | GUI | 16.2.4-3+deb9u2 | 1) | n.a. |
| BorgBackup | YES | CLI | 1.2.4-1 | ||
| BorgBackup2 | YES | CLI | 2.0.0b5-1 | ||
| vorta | YES | GUI | 0.8.10-1+deb12u1 | ||
| Cpio | CLI | 2.11+dfsg-6 | 2.13+dfsg-4 | 2.13+dfsg-7.1 | |
| dirvish | CLI | 1.2.1-1.3 | 1.2.1-2.1 | 1.2.1-2.1 | |
| duplicity | Bullseye | 0.8.17-1+b1 | 0.8.22-1+b3 | ||
| rdiff-backup | 1.2.8-7 | 2.0.5-2 | 2.2.2-1 | ||
| restic | YES | 0.14.0-1+b5 | |||
| npbackup | n.a. | ||||
| rsbackup | 3.1-3+b1 | 6.0-2+b2 | 8.0-1 | ||
| rsnapshot | Buster | CLI | 1.4.2-1 | 2) | 1.4.5-1 |
| slbackup | CLI | 0.0.12-8 | 0.0.12-13 | 0.0.12-13 | |
| snapper | CLI | 0.4.1-3 | 0.8.15-1 | 0.10.4-1 | |
| Warewulf | n.a. | CLI | |||
| Tar | CLI | 1.29b-1.1 | 1.34+dfsg-1 | 1.34+dfsg-1.2 |
Removed at 2021-12-01 due to maintainer inactivity and 9 RC bugs. See also tracker
Was removed at 2021-04-11 due to bug #986709 that tells that the upstream maintainer do not intent continue to maintain it.
Even though the acronym PC is used, it seems that this tool can be used not only for desktop or personal computers, but also for servers.
+ easy install (Debian Sqeeze, App + Webserver)
+ Webinterface easy to see
- Webinterface has bugs (not choosing of host)
- Webinterface no choice of target dir
- Webinterface no creation of host cfg
- Webinterface no comparison of files?| Version | Date | Notes |
|---|---|---|
| 0.1.4 | 2023-07-31 | Bookworm; add borgbackup, restic |
| 0.1.3 | 2022-06-12 | Add reason of package removal from Bullseye |
| 0.1.2 | 2022-06-11 | Shell -> bash, Bullseye |
| 0.1.1 | 2021-06-08 | Typos, Slbackup, Details, Dirvish, Duplicity |
| 0.1.0 | 2021-06-07 | Initial release |