-2

I'm looking for recommendations for a backup solution for our Infrastructure.
We have lots of old Hardware, which is not very fast but just laying around.
We have like 20 Servers for backup with 8-16GB RAM, AMD FX4300 to FX 8350 and many different HDDs, every Server has just a 1Gbit Ethernet nic.
At the Moment we are backing up ~80 linux servers with a rsync script directly to their associated backup server.
This Data is from our customers and the backups are not guaranteed, so we have no redundancy.

We like to change the whole concept and are looking for a solution like ceph with a good backup frontend.
I already installed a ceph cluster and the bareos frontend.
The problem here is, I couldn’t figure out, how to configure it to back up the servers to multiple endpoints (apparently this is not possible?).
This is a major prerequisite, because we have just 1Gbit/s everywhere. Backing up 80 servers to just one system is a bad idea.

Is there a solution where you have a backup manager, which manages all those backups, but is not a storage endpoint itself, instead telling every server wehre to back up?
The alternative solution would be, stick with the rsync scripts and use ceph as backend, but then the management would be as bad as before.

Simon Sutter
  • 1
  • 1
  • 1
    Given what you've said I suggest you can a local consultant in to help you work out your requirements and recommend some solutions. I don't think simply install Borg backup (which I used to use), Restic (which I now use), or another open source tool will do the job. I suggest you need to consider requirements, process, required redundency, on-site vs offsite, tape vs disk vs disk array vs cloud, and probably a whole lot of things I haven't even considered or mentioned. – Tim Mar 12 '20 at 08:40
  • 1
    That will never happen. It should be replaced as cheaply as possible and without great expense. The only one who can invest time in it is me. Everything else is too expensive. If I cannot suggest an acceptable solution, the old solution will not be replaced. Restic looks promising, maybe i'll give it a try. – Simon Sutter Mar 12 '20 at 09:17
  • Backing up many workstations to many other workstations is unusual, and probably somewhat difficult. If you could somehow pool their storage using some kind of virtualisation it might be easier? Restic is good, not sure if it will work for you. Also, what are you backing up, whole OS images, or just data from servers? – Tim Mar 12 '20 at 19:05
  • Maybe mount the ceph storage to every server an do a "local" backup is a better solution? It's just data being backed up. – Simon Sutter Mar 13 '20 at 10:16

1 Answers1

0

Maybe this will help you, take a look at:

https://borgbackup.readthedocs.io/en/stable/

We do have many servers an all of them run backups with borg. Very usefull tool.

VictorMJ
  • 108
  • 1
  • 5
  • Do you have a single borg instance where all backups go through? – Simon Sutter Mar 12 '20 at 16:03
  • I've heard about this tool, however, do not know anyone using it. That's why I did not take a risk using it. – Stuka Mar 12 '20 at 20:14
  • I used Borg for a while. It's a good tool, but for smaller backups or change sets it recreates the backup file on every backup. That is inefficient when you are uploading backups to the cloud. I replaced it with Restic which is block based and more efficient with bandwidth. The only downside of Restic is while it does deduplication it doesn't do compression yet. – Tim Mar 13 '20 at 20:01
  • Thanks for the details. – Stuka Mar 27 '20 at 20:38