views:

648

answers:

7

Duplicate: http://stackoverflow.com/questions/15208/whats-the-best-linux-backup-solution

I am about to purchase a dedicated linux server that has 160GB hard disk and will be used primarily for website hosting.

The server doesn't come with backup/restore facilities and wondered if anyone knows how / what is the best (and cheapest) method for backing up the server on a daily basis.

The server is rented so I cannot physically backup onto tape / dvd etc.

I have little linux experience so wondered if there is a open source web based interface that can perform the backups / restores etc.

+2  A: 

At work, we use BackupPC over a remote SSH login. BackupPC is an open source, disk-based backup system.

We also evaluated Amanda and a variety of other open source backup packages. None of these actually require the use of tape drives, but most of them still use a heavily tape-based model. And we didn't particularly enjoy worrying about virtual tape volumes.

BackupPC, on other other hand, requires a single large disk partition to use as storage pool, and it stores content using a chained hash of file contents. This allows it to consolidate files that are duplicated across multiple servers and multiple backup runs. So if you have multiple servers with duplicated content, BackupPC will use considerably less disk space.

If you're only trying to back up a single server, then rsync may also be an excellent choice. But BackupPC has a web-based administration interface, and it supports incremental and full backups. It can be configured to avoid backups during business hours. (However, if one of your servers hasn't been backed up for a long time, BackupPC will go ahead and grab a backup whenever it can.)

Overall, it's a very simple system to maintain. It runs without any human intervention, and it e-mails you if something goes wrong or if a server's most recent backups are getting too stale.

emk
I am sysadmin for a small company (in my spare time) and I've used BackupPC successfully. It is really easy to maintain and install.
Subtwo
+1  A: 

On a linux machine you can use RSYNC!

Or write a simple shellscript which does a SSH login, crate a .rar/.zip/.bz2... and moves it over (maybe via scp) to the backup server.

If you need a commercial product with a good support, you should have a look at the IBM Tivoli backup solutions.

Martin K.
tar + scp works well for me. If only it were that easy on Windows.
Chris
Have a look at the mingw windows tutorials! There you can use scp, tar and shellscripts well on windows.
Martin K.
+1  A: 

If it were me, I would consider two things: either Tarsnap (http://www.tarsnap.com/) or using rsync to your home machine.

The advantage of Tarsnap is that it will be easy and fast. The disadvantage, of course, is that it's not free... But it's still worth mentioning, just because it rocks.

The other option -- using rsync to copy the data back to your home machine -- is the cheaper of the two, although the first run will take a while (after that, though, it will only copy over the parts of files which have changed, so it will be much faster).

If you want more information on using rsync... It can be found easily on the web, or post here and I'll include the backup script which I use.

David Wolever
+1  A: 

Do you need to make an exact image of the server, or you just want to backup some data?

If the latter (If you just want to backup some data), you may use a utility like rsync available on most linux distributions (like Debian).

The good thing with rsync is that you're getting incremental backups, and it's algorithm doesn't need to know all the source and destination's content to run: that means small transfer bandwidth after you've done the first initial backup. All the traffic is also encrypted, if you use rsync over the internet. If you're on Debian, simply use aptitude to install the maintained version:

apt-get install rsync

If I had access to the hardware, I'd buy a raid controller and do a RAID 1 (mirror) with the existing hard drive. In this case, if one of the disk drive dies, the second one will fail over and all what you'll have to do is replace the dead drive.

Hope it helps!

Wadih M.
+1  A: 

Use Jungle Disk, and backup to a cloud, amazon s3 or mosso

Elijah Glover
I use Jungle Disk for my windows backup. Can this also be done using command line in linux?
Belliez
It says it's linux supported, install and give it a demo, or look at support docs
Elijah Glover
The new Jungle Disk Server Edition is in beta. I've been testing it on several servers and love it. It's a great solution for servers. The pricing will be low, as always from Jungle Disk. http://blog.jungledisk.com/2009/10/06/jungle-disk-server-edition-free-public-beta/
Martijn Heemels
A: 

I'm using Duplicity. It's basically rsync, except that it's encrypted, which doesn't hurt in any case. Just remember to separately back up the keys :-)

Joonas Pulakka
A: 

I use backup-manager

Backup Manager is an easy-to-use tool for generating archives.

   It  is  designed  with  simplicity  in mind for those who don’t want an
   obfuscated tool for making tarballs.

   There are different methods for building archives: tarballs,  incremen‐
   tal tarballs, mysql dumps, svn dump. You can also write your own method
   with a generic method that will execute a command and will archive  its
   output.

   Every times you run Backup Manager, you can upload archives thanks to a
   couple of upload methods: trhough FTP, SSH or even with RSYNC.

   It’s also possible to export your archives to removable media like  CDR
   or DVD.
Luis Melgratti