|
Size: 15685
Comment:
|
Size: 15738
Comment: Minor edit, incomplete line I overlooked. The article is a bit lengthy now.
|
| Deletions are marked like this. | Additions are marked like this. |
| Line 125: | Line 125: |
| The following command will overwrite drive hdb to have the same data as hda. It will wipe out all | The following command will overwrite drive hdb to have the same data as hda, this will wipe out all data previously on hdb, ensure it's what you want. |
Introduction
Backup is essential. Without it, data is entirely exposed to loss under a number of circumstances such as, drive failure or data corruption. Users often only learn this lesson after critical information they require is permanently lost. Attempting to recover from such data loss can be both time consuming and extremely troublesome. So learn from others mistakes, and ensure beforehand that you have a system in place that protects your data and suits your needs.
Backing up files can be accomplished either from a graphical user interface (GUI) or by issuing commands to the command line interface (CLI). It doesn't matter which method you employ, so long as it works for you and you do it on a regular basis. To make sure it's regular, it's best to ensure your method is scheduled so you don't need to remember to do it.
It is also important to note that there are two primary means of backup. Folder synchronization and the creation of archives.
Folder synchronization is a means to ensure two different folders at two locations hold the same data. The synchronization is usually modified by certain rules, such as directing the program to only write new files. It is most useful for multimedia files that will not benefit from compression. The reason for this is that during encoding of multimedia near maximal compression is usually achieved.
Creating an archive is mainly used for long-term storage, primarily of configuration files and documents. These will benefit greatly from the compression of gzip or bzip2 because they aren't stored in compressed formats. Since the size required on disk is much smaller than storing copies of multimedia files, users often create numerous full backups allowing them to go back and restore files from certain dates.
A third alternative is called drive imaging. Drive imaging is similar to the creation of archives, except that instead of copying just folders and files, the entire partition that is selected is backed up to an archive byte for byte. Such archives are useful, in the case of a hard drive failure, entire partitions can be restored quickly to a new hard drive and work resumed.
WARNING: Backups such as drive imaging work at a low level and if improperly used can result in severe data loss, especially with a command line program like dd. Ensure you understand what your doing!
One final note, some file systems support versioning, which allows you to see all changes to a file. This is a local backup, but will do nothing in the case of drive failure. See the appropriate section for a discussion.
Backup From a Graphical User Interface
GUIs offer a graphical interface from which a user can choose from a limited selection of options to customize operation. There are also usually configured defaults and guides that help in selecting the right options. There are a few to choose from, below you'll find several.
Simple Backup Suite
Simple Backup Suite is a program designed for desktop users. It's purpose is succinctly summarized by it's title, to be a simple backup suite. It is easily installed from the repositories and comes with defaults already loaded and configured. From there you can customize what directories it backs up, what it excludes, schedule it to certain times and manage old archives. This suite creates archives, keep that in mind during use.
For a thorough guide to installation and general use, please see BackupYourSystem/SimpleBackupSuite.
grsync
The program grsync is simply a GUI front end for the CLI utility rsync. It presents a clean yet simple interface for controlling the synchronization between two folders. It also supports multiple sessions (otherwise called profiles) and many of the commonly used basic and advanced options for modifying the synchronization.
For more information see the rsync page.
Backup From the Command Line
CLI's primary advantage over a GUI is that it exposes directly to you many more options. With that however comes the obligation of knowing exactly before hand what you want to do. There are no defaults with commands it's entirely up to the user to set the desired options. To better acquaint yourself beforehand, it is advisable that you read UsingTheTerminal.
Creating an Archive
When backing up from the command line TAR is often used. The TAR command creates an archive file, which can then be further compressed by gzip or bzip2 to decrease the size it occupies on the disk.
Depending on what you are backing up superuser access might be needed. For instance, if you want to backup your configuration files in /etc. For an understanding of why this is needed, see FilePermissions.
To begin open Applications -> Accessories -> Terminal and type the following command and return.
sudo -s -H
This will give you root access for the duration of your terminal session. Take care with the commands typed from here on out, you have access to everything on your disk.
For this example, we will change directory to the root, this is where the archive file will be created. You may decide to use another directory, including external drives or other removable media. That is entirely up to you.
cd /
The following command will backup your computer's important files, to understand the specifics please see the TAR page below it.
tar cvpzf my_backup.tar.gz --exclude=/proc --exclude=/lost+found –exclude=/my_backup.tar.gz --exclude=/mnt --exclude=/sys /
A full explanation of this command, including how to restore and how to operate over a network among other things can be found at BackupYourSystem/TAR Read before using the command.
rsync
Rsync updates the copies the files that have changed and even then only transfers the parts of those files that have changed. That is useful for saving bandwidth when backing up over the network. For safety, transfer between two machines is done via SSH. Rsync is especially good for backing up home directories.
The command for transferring to a remote machine is:
sudo rsync --delete -azvv -e ssh /home remoteuser@remotehost.remotedomain:./backupdirectory
-z compresses the data --delete deletes files that don't exist on the system being backed up. Maybe you want this, maybe not. -a preserves the date and times of the files (same as -t), descends recursively into all directories (same as -r), copies symlinks as symlinks (same as -l), preserves file permissions (same as -p), preserves groups (same as -g), preserves file ownership (same as -o), and preserves devices as devices (same as -D). -vv increases the verbosity of the reporting process
rsync has several graphical frontends: grsync, Backup Monitor, QSync, Zynk, rsyncbackup, TKsync
For more options, there is a separate page about rsync. As a companion, the article on SSH might also be worth reading.
Backup Destination on a Remote Machine
Simple Backup Suite
BackupYourSystem/SimpleBackupSuite
Duplicity
BackupPC
Drive Imaging
Drive imaging (also called disc imaging and partition imaging) is very useful, but only in specific cases. In general, it isn't an operation a user would run weekly or daily, rsync and archive creation are much better for such backups. Drive imaging is better used as a long term archive of a partition. Say for instance a user's home machine is set up to dual boot between XP and Ubuntu, any number of things could corrupt his XP or Ubuntu partitions and necessitate a re-installation of either. With a drive image, a user could simply format the partition and copy the data from the archive into the partition and be back up and running quickly.
Drive imaging therefore, like all the other means described is complementary to a user's backup strategy and should be considered on it's merits.
Graphical Programs for Drive Imaging
Two good graphic textual-graphic programs are partimage and clonezilla. They can be found at the below sites.
Creating Disc Images Using dd
WARNING: Read carefully and don't make typos!
dd is an old command line program used for low level copying of files. It is a legacy program from UNIX, as such it does not behave or configure like traditional programs in the GNU suite. Caution is advised, using this command with the wrong operators can cause serious data loss.
In the following commands, if and of are short for input file and output file respectively. As the name implies, the input is the partition being backed up and output is the archive created. Last important note, you need to know what partition you want to backup and what it's label is. To find out, open a terminal and use the following command:
sudo fdisk -l
The output of the command will list each hard drive installed, and all the partitions on it. It is usually possible to tell which is which by looking at the filesystem installed (ntfs for Windows, ext3/4 for linux) and the size it occupies on the disk. Your responsible for knowing which partition is to be backed up. In the following examples replace /dev/hda1 with the name of the partition listed under device boot in the output.
The following example will create a drive image of the partition hda1.
dd if=/dev/hda1 of=/home/hda1.bin
This command creates the drive image and then passes it through gzip for compression. Bzip2 could be substituted instead of gzip.
dd if=/dev/hda1 | gzip > /home/hda1.bin.gz
Back up the MBR of the hard drive labelled hda.
dd if=/dev/hda of=/home/hda.boot.mbr bs=512 count=1
Restore MBR (from a Live CD) to hda.
dd if=/mnt/hda1/home/hda.boot.mbr of=/dev/hda bs=512 count=1
The following command will overwrite drive hdb to have the same data as hda, this will wipe out all data previously on hdb, ensure it's what you want.
dd if=/dev/hda of=/dev/hdb conv=noerror,sync bs=4k
The command:
dd -if /dev/hda1 > partitionimage.dd
will backup "/dev/hda1", the first partition on hda1. A whole drive (including the MBR) could be backed up using just /dev/hda as the input "file". Restoring is done by:
dd -if partitionimage.dd -of /dev/hda1
If you have a complete new harddrive and want to restore the backup (or copy your old system to the new drive). First, the new drive has to be bigger or exactly the same size as the old one.
First go superuser and switch to runlevel 1 so that you can fumble around with the harddisk without other services interfering:
sudo su init 1
restore either the whole disk to the new drive or one partition (depending on how you made the backup):
dd -if partitionimage.dd -of /dev/hda1
If you restored the whole drive (/dev/hda), the system will not automatically create the devices (/dev/hda1, /dev/hda2) if you just restored the whole drive. If you know how to make the devices show up without reboot, write it here, otherwise this is a good moment to reboot.
If you restored the system to a new drive, and your device names changed (for example from /dev/hda to /dev/sda) then you must adapt the bootloader and the mount points. While still on runlevel 1, edit these files:
/boot/grub/menu.list /etc/fstab
Update your drive mappings to the new devices there. Grub may be setup with numberish IDs for your drives, then you need the command blkid to read the new identifiers of your disk.
After your system is able to boot and runs again, you can resize your partitions to fill the rest of the new harddisk (if you want that) as described here:
You can use the "loopback device" to mount this partition image, making it easy to read the old version of files and still keep your current version untouched. Mounting disks with Linux's loopback device by Jason Boxman
Versioning File Systems
These will be very useful when they mature. For example, ext3cow file system states it, "provides a time-shifting interface that allows a real-time and continuous view of the past. This allows users to access their file system as it appeared at any point in time." This feature is automatic -- nothing (meaning no backups, manual copies, etc.) need be done to have this benefit when using a copy-on-write file system.
They let you see every change to a file over it's life span, but offer no protection against a hard drive that fails. These projects are therefore complimentary to the backups described above, which offer snapshots or synchronization that can be stored on other hard drives or even other locations on a network.
Versioning tools such as Subversion could also be useful while we wait for versioning file systems to become mature and fully supported.
Related Links
You might also want to check out these backup programs which will help you to make automated backups of your system:
nice.
Other wiki discussing various Linux backup techniques:
See the full discussion in these threads on the Ubuntu forums:
http://www.ubuntuforums.org/showthread.php?t=35087 Howto: Backup and restore your system! (very long thread; but all the good bits have (hopefully) already been summarized on this page.)
"Backup and Restore Your Ubuntu System using Sbackup" has nice step-by-step instructions (with screenshots) of using sbackup.
A personal blog listing many open source backup solutions: List of Free Open Source Linux and Ubuntu Backup Software
Early Stage Ideas
In addition to that already discussed, there are projects still in development that you may wish to keep appraised of or even participate in.