Is there a mechanism for smoothing disk I/O in a similar fashion using nice to smooth CPU usage? I often find that certain disk-intensive tasks put the system into high wait-state load averages, and system performance as a whole suffers.
Nicing the process doesn't really help, as the system won't release anything in a wait state. The recent Debian scheduler defaults to CFQ; is this optimal? What other things can I do to keep the system from being unresponsive during heavy I/O? -- Unabashedly littering the information superhighway with detritus like this for over 15 years now. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
