block: update documentation for deadline fifo_batch tunable
Update the description of fifo_batch to match the current implementation, and include a description of how to tune it. Signed-off-by: Aaron Carroll <aaronc@gelato.unsw.edu.au> Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
This commit is contained in:
Родитель
4fb72f7646
Коммит
6a421c1dc9
|
@ -30,12 +30,18 @@ write_expire (in ms)
|
||||||
Similar to read_expire mentioned above, but for writes.
|
Similar to read_expire mentioned above, but for writes.
|
||||||
|
|
||||||
|
|
||||||
fifo_batch
|
fifo_batch (number of requests)
|
||||||
----------
|
----------
|
||||||
|
|
||||||
When a read request expires its deadline, we must move some requests from
|
Requests are grouped into ``batches'' of a particular data direction (read or
|
||||||
the sorted io scheduler list to the block device dispatch queue. fifo_batch
|
write) which are serviced in increasing sector order. To limit extra seeking,
|
||||||
controls how many requests we move.
|
deadline expiries are only checked between batches. fifo_batch controls the
|
||||||
|
maximum number of requests per batch.
|
||||||
|
|
||||||
|
This parameter tunes the balance between per-request latency and aggregate
|
||||||
|
throughput. When low latency is the primary concern, smaller is better (where
|
||||||
|
a value of 1 yields first-come first-served behaviour). Increasing fifo_batch
|
||||||
|
generally improves throughput, at the cost of latency variation.
|
||||||
|
|
||||||
|
|
||||||
writes_starved (number of dispatches)
|
writes_starved (number of dispatches)
|
||||||
|
|
Загрузка…
Ссылка в новой задаче