Hello, On Wed, 21 Apr 2021 at 13:29, <[email protected]> wrote:
> Hi, > > I'm using a drbd 9.13 on CentOS 8 with pacemaker for a mysql-cluster. This > cluster is running on a kvm/qemu cluster. You are probably referring to the drbd-utils version as the most current drbd kernel module version is drbd-9.0.28. The virtual harddisks for the virtual mysql-nodes are qcow2 devices. > Is this really a good idea to use qcow2 as base for drbd ? Since drbd is a block device I would not use qcow2 on top of it, mainly for performance purposes.I'd rather deploy the VMs directly onto it (raw). I can see that the qcow2-devices have grown larger as the drives of the > virtual machines really are right now. Is this normal ? Do you mean that the size of the qcow2 on the host has grown up larger than the actual "used" size withing the guest? If yes, then I think that's a case of correctly configuring the trim or discard support both on the guest and the host (not related to drbd). Yanni
_______________________________________________ Star us on GITHUB: https://github.com/LINBIT drbd-user mailing list [email protected] https://lists.linbit.com/mailman/listinfo/drbd-user
