On Centos server I am seeing the major-faults count go up drastically during peak load. We have disabled swap and we rarely modify content of disk (we read data off network, pr