pool question

[Frank Cherry]

Hello there,

I have running a Proxmox cluster which backups the VMs and CTs daily on a disk array with a history of 30 days.
The only changing data inside the virtual machines are error logs (unimportant if restore case) and configurations (important).
Configuration changes are done 1 to 3 times per month. Important user data are stored outside. So important for backup is the VM with it's config as a working unit.

My plan is to set up a job without scheduler for each VM - so I have it run manually when I have made config changes on a VM or had run an update on it. It should be for each job a full backup - the compressed backup file of a CT has around 3 GB.
The backup should be store on two different tapes (LTO-7)

So calculated in append mode, with 6000 GB space on tape, I would have space for up to 2000 full backups per tape.

Is there a way to do it within a pool definition, e.g. make a copy on a 2nd volume or the other way for me would be:

1. Making a spooling job to local disk
2. followed by a copy job to tape volume
3. followed by a move job to 2nd tape volume

When using the last scenario, can I label both volumes identically and it would be no matter which tape I use in the case of restoring?

I do own a single tape drive only, no autoloader. So manually work is needed.

Thanks, all the best, Frank

[Brock Palen]

I don’t trust snapshots for backup, though I use them heavily at the day job on many PB of data. This does not protect against bad firmware, site disaster, admin error, etc. You're choice how much you value the data. If it’s personal then this is probably fine. Otherwise defense in depth and use both.

I personally do not like manual jobs, humans tend to forget and opens options for error. I would encourage some level of automatic even if just a scheduled monthly full, in addition to your manual runs. If you can really fit 6000 fulls/tape an LTO7 tape is $55 right now. Sounds like cheap insurance.

As for two tapes you will need multiple drives to keep it easy. I only have 1 drive for my home backups and tried doing what I will sketch out but never quite got it working reliably.

If you have 2 drives I would not bother with the disk spool job. Just use a spool as part of your job to tape off an SSD or fast array so you can stream to the drive.
https://docs.bareos.org/TasksAndConcepts/DataSpooling.html

Then run a nightly copy job to copy any new jobs on the one pool (in one drive) to a second pool
https://docs.bareos.org/TasksAndConcepts/MigrationAndCopy.html#

I personally tried doing a write to disk (fast restores of latest) a copy to tape, and then a migration to a different pool (single drive with auto loader) It always gave me trouble so I gave up on it and instead take an Archive job on a cheap LTO4 drive I picked up for $99 on ebay every month. I have to swap 3 tapes by hand but it’s helps.

Be be clear the issues I had were not so much with shuffling around the data so much. I’m using Always Incremental for road warriors so we can’t do full backups except when we provision a new system to them or we ship out a RPi in a box to do a full backup and sneaker net it. Bareos really does not handle multiple copies with AI well, related to this bug: https://bugs.bareos.org/view.php?id=778

Using the archive job is a work around but not managed daily in my current setup.


Brock Palen
bro...@mlds-networks.com
www.mlds-networks.com
Websites, Linux, Hosting, Joomla, Consulting

> --
> You received this message because you are subscribed to the Google Groups "bareos-users" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to bareos-users...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/bareos-users/6bd90c53-9331-466d-a324-5c8a2265c2afn%40googlegroups.com.