How to implement incremental + full backup linux?

Question

css

I don’t have enough experience in this matter :( I
need advice on this ...

1. I thought about how easier, and most importantly better :) to implement backup of a linux server (LVM + RAID), to a remote network drive, and then a number of questions arise :
1.1 Many utilities come to mind, but which one will be easier to cope with and at the same time not run into a rake in the future, only a practitioner can answer this question ... Of course, I would not want to fence the garden with BareOS, while I tend to rsync and Clonezilla;
1.2 I would also like to monitor the functioning of the work through Zabbix, but have not yet figured out which side to approach (Zabbix itself is already installed);
1.3 Accordingly, I would like to create backups without shutting down and restarting the server.

2. I would like to master the incremental backup:
2.1 What tools would you recommend for this?;
2.2 In the event of a complete server failure, how difficult is it to assemble a system for deployment from these incremental copies? Or will you have to deploy the system from the beginning, and then on top of it is a piece of incremental backup?
2.3 Maybe just once a week (I don't need more yet), copy an LVM snapshot? It will then be necessary to re-create it regularly and monitor it with Zabbix so that this business works ...

3. All this business will probably be managed by Kron. And what else? :)

4. In Zabiks, you will probably have to come up with manual checks for the operation of creating backups, and at different stages.
4.1 Of course, in this case it is easy to make a mistake and miss something, which begs the question: Are there any ready-made solutions?

Answer 1

One common container, inside 3 divs with content. Container display: flex, desired width, justify-content: space-between.

Answer 2

Make TK, we'll finish it :)

Answer 3

I am using borgbackup . It works for me everywhere and I even made my own creation for docker and kubernetes .
He has a distinctive feature that he always makes a full backup, but repeating blocks simply refer to existing ones. As a result, when restoring, you do not need to first restore fullbackup first, and then each increment. Well, out of the box, he also does deduplication and archiving, plus everything is encrypted.
For example, I (at work) have a 2.8T file dump and a 60-day backup takes 1.4T.

Answer 4

How many numbers and bukaf.
tar+cron and rsync will save the father of Russian democracy. The first one is for increment and full, the second one is for full.

Answer 5

Bacula.

Answer 6

Correctly written above. Technically, tar+cron solves this problem. And usually the simplest solution is the most reliable. If you "want to track" - then this is another task. And it is not directly related to the backup. That is, the question consists of two independent parts. Actually backup. And some kind of monitoring that should ... monitor something.

Answer 7

Rsync is ideal for managing small to medium size backups. An incremental backup based on hard links will allow you:
- on the one hand - to have completely independent full backups, with the ability to delete unnecessary and elementary recovery of any file at any time
- on the other hand, disk space optimization, because. in reality, only different files will be stored on the disk - the rest will be connected by hard links and will not take up additional space.
So, if you're comfortable with file-level storage and don't have large, frequently changing binaries (such as database files), then rsync (and its many add-on utilities) is a great option!
However, you wrote that you want to do LVM snapshots remotely. That is, you need the ability to transfer binary blobs (or even better - only the difference) remotely, optimally store and, judging by the fact that the backup server does not belong to you, in a good way, also encrypt. All this is in, already recommended to you, BorgBackup. This is such a git repository for binary files with deduplication and encryption. Unfortunately, I can’t say anything about his work with win-hosts - I haven’t come across it.
As for monitoring: it makes sense to either immediately look for a backup system with monitoring out of the box, or choose a convenient tool and add monitoring to it. In your case, I would choose the second option.

Answer 8

What's on this server? It may be easier to drive the entire configuration into Ansible, namely, changing data (database, or whatever) back up with specialized tools.

Answer 9

I use the backupninja add-on for rdiff-backup and more.

Answer 10

Veeam Agent Free.
We use both on Win-machines, and on Linux. I had to restore Windows more than once, everything works. Linux machines never had to.

Answer 11

I wonder why no one walked through Bacula. A slightly tricky config, but if you sharpen everything to your needs, the thing is very worthy. Also lightweight rdiff-backup , well, they correctly write rsync itself above. If you have the desire and time, you can write on its basis a wagon and a cart of scripts for backing up the system.

Answer 12

Does anyone remember such a thing: FSBackup ? :)
https://www.opennet.ru/dev/fsbackup/