Disk Space Allocation Concerns with Longhorn

Hi all,
I’ve setup CrateDB using the crate-operator on a K3S cluster using Longhorn storage.

I have some concerns about the disk allocation.

I defined a debug disk of 4GB and a data disk of 16GB. This on a 3 node CrateDB cluster means 60GB disk space on 6 PV.

But I’m using a 3 node Longorn cluster so any PV is replicated 3 times.

This means that for 16GB of data I need 180GB of disk space.

Is this expected or I’m doing something wrong?

Disclaimer: I don’t have experience with Longhorn.

In a 3 node setup, with 20GB for each node, 60GB will be used.

If Longhorn replicates the volumes you set up, three times, I guess it makes sense it ends up being 180GB.

CrateDB already tries very hard to maintain availability and durability; read more here, may I know more about your use case or your expectations on disk usage?

1 Like