Sequential creation of Volumes when scaling MachineDeployment #2431
Description
In our setup when provisioning a VM on OpenStack, a volume is first created with the image of the OS, from which the VM later boots.
When scaling MachineDeployment, the desired number of OpenstackMachines objects is created at once but regardless of parameters (we use default values):
--openstackcluster-concurrency int Number of OpenStackClusters to process simultaneously (default 10)
--openstackmachine-concurrency int Number of OpenStackMachines to process simultaneously (default 10)the volumes are created sequentially and in turn, the virtual machines are provisioned sequentially as well.
The effect is a long time when scaling from a small to a large number of MachineDeployment replicas.
Is this an intended behaviour of CAPO? Can the concurrency of volume creation be increased?