Am now in the process of “thrash testing” Kopia, and one of the problems I faced was drivepool compatibility through an SMB connection. It seems the pool could not keep up with Kopia (writes). Accordingly, I set that project aside for now, and may attempt to setup Kopia for direct disk access to a drive pool. If it works, would then consider using multiple drivepool targets enabling off site backups and ransomware protection.
If any of you are using drivepool like this - would appreciate your insights and would be grateful to profit from your failures
The advantage here is that your Kopia repository can be easily expanded. You aren’t limited to a single large drive to host your backups.
What makes you to say it? Is it slow? If you store data on system with very low IOPS you should try to lower kopia parallelism but it will be slow regardless. Kopia operates on many small files which often is the worst case scenario for mechanical disks.
Drivepool does take a hit on writes when used with redundant disks (only way I’ve ever used it). Write speed may not be an issue with a simple aggregated virtual disk - but I’ve never tried it.
The only way to know for sure is to spend some money and try it. What I was hoping to avoid
It’s possible that combining Kopia with drivepool will give it “legs” …because there’s really no limit to teh number of aggregated disks (means you could keep all backups deduped because it’s all in one place). Now how all this behaves when a single disk fails within the array is another interesting question. Best case, there would be no data loss after you rep[laced the disk and it would rebuild rapidly? Worst case, you would lose all your snapshots AND your data. Requireing a complete new backup (the the prospect of more disk failures hanging over your head).
Currently, I use Kopia to backup what is essentially a RAID 1 array. But I might like to get more adventurous.
