pf » Howto Backup your Mac incrementally over SSH

Howto Backup your Mac incrementally over SSH

applelinux

Do you have access to a shell account on a unix server with some spare space? If so it's pretty easy to incrementally backup your files securely with SSH.

I titled this entry Howto Backup your Mac incrementally over SSH but this technique can also be used to backup any computer that can run rsync and ssh. They are already installed on Mac OS X, and most linux / unix servers.

Step 1 - Create a folder to store your backups on the remote server

mkdir backup

Make sure that your SSH user has permission to write to this directory.

Step 2 - Setup automatic authentication Optional

This step allows the backups to run without prompting you for a password when it runs. You can omit this step but you will have to type in your ssh password each you run backup.

I wrote an article called Setting up public key authentication over SSH that will guide you through this step.

If you own the server you might also want to create a user specifically for this process.

Step 3 - Use rsync to backup files incrementally

rsync -e "ssh" -rca --delete-after ~/test/ pete@myserver.example.com:backup

Now lets break it down a bit:

  • rsync - this syncs the local directory to with the server directory.
  • -e "ssh" - this tells rsync to use ssh if your want to pass in other ssh options such as port you can do that in the quotes: -e "ssh -p 12345"
  • -rca recursive, checksum, and archive
  • --delete-after - this will delete files on the server if you delete them locally.
  • ~/test/ - I am backing up / syncing the test directory inside my home directory on my mac.
  • pete@myserver.example.com:backup - my ssh username is pete, my remote ssh server hostname is myserver.example.com, and I am backing up into the directory ~pete/backup.

Excluding directories

Sometimes you might want to exclude a directory from being backed up, perhaps your Music directory since that is already backed up on your ipod.

rsync -e "ssh" -rca --delete-after --exclude=Music --delete-excluded ~/test/ pete@myserver.example.com:backup

Step 4 - Schedule it with cron Optional

Now lets create a cron job (scheduled task) to run this script every day. First make a new file called backup.sh in your home directory.

#!/bin/sh
rsync -e "ssh" -rca --delete-after ~/test/ pete@myserver.example.com:backup

Now sure make the file is executable: chmod ug+x backup.sh

Next type crontab -e this will open up your scheduled tasks list, in most cases it will open an empty file. Add the following to the file:

15 11 * * * yourusername ~yourusername/backup.sh

That's will schedule the script to run at 11:15am. If you don't want to do this step you can simply run the command whenever you want to run a backup.

DISCLAIMER You should test everything before performing the backups on your live data. If you would like to use the scripts above you must acknowledge that I'm not responsible for any loss of data.

I'd love to hear your feedback on this, please comment below. I'm use this technique to backup my two web servers on to each other.



Related Entries
22 people found this page useful, what do you think?

Trackback Address: 549/5525DF6513765A55A1D7347BC5D3D540
On 03/10/2006 at 8:33:28 AM MST Doug Laakso wrote:
1
Pete, You may be interested in the FreeNAS project at http://www.freenas.org/. This OS can turn an old computer into a File Server with everything you have described here.

On 03/10/2006 at 9:52:41 AM MST Jon Gales wrote:
2
Thanks Pete, I have it working like a champ. A dedicated server of mine is backing itself up to a shared server account I have. They both have fast connections so it's going really quickly.

On 03/21/2006 at 7:15:07 AM MST K Jim wrote:
3
I tried this approach, but...the resource forks are not preserved. Of course, if your data doesn't require this, it's a great solution.

On 06/14/2006 at 5:35:44 PM MDT Jok wrote:
4
be careful. this will only backup the main stream of each file. applications for example, that you would drop in folders as part of the backup wouldn't actually get backedup properly (they would show up as folder, loosing all ability to run...)

this is good to backup datafiles (that are single-streamed), but not all files are.

make sure that the copies of the files you have are readable on the backup media...

On 08/09/2006 at 9:34:07 AM MDT Nav wrote:
5
I didnt quite understand the last comment if someone can put in little more layman terms i would appreciate.

On 01/09/2007 at 12:38:31 PM MST speedeep wrote:
6
Beware "--delete-after". I run two scheduled cron jobs, one S,M,Tu,W,Th,F without the --delete-after option, then a separate Saturday job with the --delete-after option. If you run this automatically always using the --delete-after switch, you will find that the file you accidentally deleted yesterday was faithfully removed from your backup copy as well.

On 03/06/2007 at 11:54:07 PM MST Ronnie wrote:
7
try using backup software that supports it like www.disksave.com

On 02/25/2008 at 8:05:56 AM MST Daff wrote:
8
Hi I want to recommend you very useful rapidshare search http://loadingvault.com. You can find there a lot of new movies, games and music. Enjoy it!

On 03/11/2008 at 10:31:29 AM MST tatianahunt wrote:
9
It very important. It almost started to work. Fileshunt.com has incredible speed of searching rapidshare links in the internet. http://fileshunt.com database includes all rapidshare links.

On 03/26/2008 at 7:58:47 PM MST Marty wrote:
10
Here's a script I wrote for backups years ago.

Change the following variables:

- RSYNCUSER=unix username on remote host - EMAIL=your email address - BACKUPSERVER=your remote hostname - KEY=path to your ssh key (passwordless login)

#!/bin/bash LABEL="BACKUP-`date +%Y%m%d`" SUMMARY="/var/tmp/summary" TIME=`date +%k:%M:%S` BACKUPDATE=`date "+%A %d/%m/%Y"` EMAIL=your@email.address RSYNCUSER=your_rsync_username BACKUPSERVER=backup.server KEY=/root/keys/backup.server.key

check_server() { ssh -i $KEY $RSYNCUSER@$BACKUPSERVER exit >> /dev/null 2>&1 if [ $? -ne 0 ]; then printf "Backup did not run. The backup server was unavailable.\n" >> $SUMMARY cat $SUMMARY | mail -s "Backup summary for $BACKUPDATE" $EMAIL exit 1 fi }

run_backup() { printf "Filesystem backup performed on `date +%m/%d/%Y`.\n\n" >> $SUMMARY printf "\nSession summary:\n" >> $SUMMARY printf "\n\t\t`date +%k:%M:%S`\t- Session started\n" >> $SUMMARY printf "\n\t\t`date +%k:%M:%S`\t- Processing backup\n" >> $SUMMARY rsync -ao --stats -e "ssh -i $KEY" / $RSYNCUSER@$BACKUPSERVER:/ --exclude='/proc*' --exclude='/dev*' > /var/log/backup.log 2>&1 if [ $? -eq 0 ]; then printf "\n\t\t`date +%k:%M:%S`\t- Session completed\n\n" >> $SUMMARY else printf "\n\t\t`date +%k:%M:%S`\t- Session aborted\n\n" >> $SUMMARY fi }

check_server run_backup cat $SUMMARY | mail -s "Backup summary for $BACKUPDATE" $EMAIL exit 0

On 03/31/2008 at 1:48:00 PM MST iceman wrote:
11
This will work fine for resource forks and other metadata and extended attributes if you have at least 10.4 and add the "-E" switch to the rsync command.




  



Spell Checker by Foundeo





Subscribe to my RSS Feed: solosub RSS
Tags