On 11/10/2016 08:04 AM, Hannes Reinecke wrote:
this really feels like a follow-up to the discussion we've had in Santa
Fe, but finally I'm able to substantiate it with some numbers.
Hello Hannes,
Since for the sequential read test the difference between the 50th
percentile sequential read completion time (2.7 ms) and the average
completion time (52 ms) is large I assume that you ran the ssd-test.fio
job on top of a hard disk or hard disk array? How many hardware queues
were active during your test? If multiple hardware queues were active,
can you repeat your test with a single hardware queue? The plan for
adding I/O scheduling support to blk-mq is - at least initially - only
to perform I/O scheduling if there is only a single hardware queue.
Bart.
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
