We recently installed a Hadoop cluster using Dell machines such as R750/R760, etc. Each machine has 14 disks, with each disk having a size of 16TB , and Hadoop cluster installed on RHEL 8.6 version
For some unclear reason, we found that the disk performance on this cluster is quite unique, with disk utilization spiking to 100% or fluctuating in a way that jumps to 100%. This is in contrast to another cluster where the disk utilization remains very low, between 10% to 20%.
So, I would like to ask a basic question about the disks: Is it possible that this disk utilization behavior is due to inconsistencies between the disks and the controller, the firmware version of the controller, or the disks and the hardware of the machine?
