It is possible that while a disk request for a particular cylinder is being serviced, another request for that cylinder will arrive. Some disk scheduling disciplines would service this new request immediately after processing the current request.Other disciplines preclude the servicing of the new request until the return sweep of the disk arm.What dangerous situation might occur in a disk scheduling discipline that allowed immediate servicing of a new request for the same cylinder as the previous request?

What will be an ideal response?

Indefinite postponement might occur if a steady stream of requests arrived for that
cylinder. One way to avoid this would be to service requests up to some fixed limit and save
any additional requests for the return sweep.