How to make backups?

Question

Amazon Web Services

Finally, I decided that I needed to change something in my life and decided to “start making backups” ©.
I plan to use some kind of Amazon service as a target storage (if there are objective reasons why Amazon is not very good, I will be glad to hear).
But how exactly do you make backups? It is clear that everything can be scripted. There, in theory, there is nothing complicated, but I want some beautiful ready-made solution.
The task is to backup several (tens) Unix servers, database servers, windows servers and workstations. The main condition is no extra software on the machines where the backup will be taken from. The same conditions - the system must support versioning for reasons of space saving. Those. if nothing has changed, we don’t backup anything. if it has changed, we only backup what has changed.

Answer 1

Well, since every sandpiper praises his swamp, then look at rsnapshot.
Both versioning and space savings are very impressive, and data encryption during transmission (through the ssh tunnel), and everything is configured very simply, and you can even screw your own scripts. Yes, and it works flawlessly.

Answer 2

I haven't implemented it myself yet. But the idea is good. Make backups in the form of deb-packages. Those. Script to collect the package, upload it somewhere for storage and that's it. Of the benefits - a well-assembled and configured package can be easily installed later. For, as they recently said in Habré: any fool can make backups, but how to restore the backup later is the main task.

Answer 3

I used duplicity. Available in the debian repositories. Able to receive data via sftp, ftp, i.e. it is enough to put it on one machine, and then configure access to other machines. Makes incremental backups, you can configure all time intervals. Able to use Amazon S3. You can roll back all files or selected files to a state corresponding to any day (if the backup is done every day). Supports data encryption via gpg.

Answer 4

At one time, when I first rented a server from Hetzener, I saw a description of Tartarus on their Wiki .
In fact, this is an add-on (bash scripts) for tar, it can work via (S)FTP, locally, emulationally. Accordingly, it can also make incremental patches.

Answer 5

I am using crashplan. There is versioning. No volume limit (yet).
The only BUT is that their client utility should be installed (works on java)

Answer 6

There is a chic gem - backup github.com/meskyanichi/backup/wiki
The configuration is written on your DSL, where it is not enough, you can use Ruby, backup databases, files, etc. on clouds, on ftp and just in a folder even. Try it, it's really very simple.

Answer 7

Look towards Bacula