Following on from the post on backup which somewhat obscurely was called Backup, backup, backup with extra backup this post looks at how the backup works
Apple has Time Machine
Set and forget backup software with a really fast restore.
Works locally
Microsoft has backup
- Comes free with Windows
- Creates snapshots
- Is encrypted
- backs up locally
Linux has lots of ways to backup.
On the desktop with a nice mouse environment there are many, many programs both commercial and open-source that backup locally and remote.
Our servers don’t waste RAM and CPU on that sort of thing when everything runs from the terminal
Rsync. The unsung hero of the Internet
The rsync utility uses an algorithm invented by Australian computer programmer Andrew Tridgell for efficiently transmitting a structure (such as a file) across a communications link when the receiving computer already has a similar, but not identical, version of the same structure.
Wikipedia
This simple idea allows vast changing archives of data to be maintained as an identical mirror on the other side of the planet or on your laptop. After the initial sync of the data between systems only the changes are transferred:
- New files
- Changed files
- Deleted files
When rsync detects a changed file at the sending side it will send the changes not the whole file. This is the real power of rsync. There are many ways to use rsync but this method is the most widely used to mirror big data across the Internet
How backup is done over even over low bandwidth connections
I made a five thousand word Word file a few days ago. This was transferred to the mirror as a new file.
Today I realized that the out of five thousand words I had made a typo. I changed the an to an and and closed the file on my local server.
The backup runs and finds the change
The backup system detected the change in the file and after comparing the documents sends only the differences. In this case the d.
Now that the files are identical again rsync will only compare the files until the next change
What about if you want a backup of the file you had last week?
We use magic to give apparent full backups with very little space
Without going into horrendous complex Linux commands and the history of hard links we can do what look like full system backups using only tiny amounts of data for each file.
The file is sent as before..
The new file is now identical on both servers.
Oh dear. You trashed your awesome report and did not see this until after the file had been transferred to the mirror server
There are now two copies of your report. The file name has not changed but the contents are different.
Recovery of data and business continuity
You will notice that there is no mention of backup sets, recovery software or special permissions
- The software preserves on all files and directories
- File access and modification times are preserved.
- User access permissions
- Identical structure for each backup day, week or month
Find out more about backup and recovery options for:
- Recovering Windows Servers
- Backing up Windows Networks