Handling large disk volumes

[Faris Raouf]

I'm wondering if anybody can offer me any advice on handling large VM disk volumes
We only use RAID10 local storage, no ZFS, no Ceph. We copy backups to a second node and also to S3.

I have a customer with a VM that has a 500GB disk in lvm-thin. He's used almost all his allocation.

Backing it up via Proxmox only takes an hour. But the backup is 450Gb. We then copy that to a second local node, and then send it to S3. Each step takes an hour or more.

And now the customer wants more disk space. He's going to end up with 1TB of disk usage. And no doubt eventually 2TB.

I don't know how to handle these huge volumes. They seem unreasonably large. All he's storing is old email that nobody looks at. It drives me nuts.

Would giving the customer a second 500GB volume be better than one 1TB volume? I just feel that 500GB is about the limit for a reasonable size single file. 1TB just seems unreasonably large, especially when total storage is 4TB.

I'm thinking of adding a pair of additional 2TB drives just to store his volume(s) and his local backups. That would at least keep him happy until he ends up needing even more space.

 

[Alwin]

This sounds like, the image has not gone through a discard on empty blocks. If possible, you or the customer could issue a fstrim and free up unused blocks.
http://man7.org/linux/man-pages/man8/fstrim.8.html

The vzdump generates a big backup file in any case, as it packs all disks into the vma file. As long as you are able to restore such a backup, the size shouldn't matter. Besides the time it takes and the space it uses.

 

[Nemesiz]

Change the strategy of backup. Use bareos inside VM to do increment backups.