Blogbench RAID benchmarks

[Freddie Cash]

On Thu, Jul 21, 2011 at 3:35 PM, Freddie Cash <fjwcash at gmail.com> wrote:
On Mon, Jul 18, 2011 at 7:06 PM, Matthew Dillon <dillon at apollo.backplane.com> wrote:

 Â  Â Ok, well this is interesting. Â Basically it comes down to whether we
 Â  Â want to starve read operations or whether we want to starve write
 Â  Â operations.

 Â  Â The FreeBSD results starve read operations, while the DragonFly results
 Â  Â starve write operations. Â That's the entirety of the difference between
 Â  Â the two tests.
Would using the disk scheduler's in FBSD/DFly help with this at all?FreeBSD includes a geom_sched class for enabling pluggable disk scheduler's (currently only round-robin algorithm is implemented).  http://info.iet.unipi.it/~luigi/geom_sched/

Page 39 of the presentation on GEOM_SCHED shows the following, indicating that it should make a big difference in the blogbench results (note the second result with greedy read and write):
Some preliminary results on scheduler’s performance in some easycases (the focus here is on the framework).Measurement is using multiple dd instances on a filesystems, allspeeds in MiB/s.two greedy readers, throughput improvement
NORMAL: 6.8 + 6.8 ; GSCHED RR: 27.0 + 27.0one greedy reader, one greedy writer, capture effectNORMAL: R: 0.234 W:72.3 ; GSCHED RR: R:12.0 W:40.0multiple greedy writers, only small loss of througputNORMAL: 16+16; RR: 15.5 + 15.5
one sequential reader, one random reader (fio)NORMAL: Seq: 4. 2 Rand: 4.2; RR: Seq: 30 Rand: 4.4 
And I believe DFly has dsched? 

 Â Â  This is all with swapcache turned off. Â The only way to test in a
 Â  Â fair manner with swapcache turned on (with a SSD) is if the FreeBSD
 Â  Â test used a similar setup w/ZFS.
ZFS includes it's own disk scheduler, so geom_sched wouldn't help in that case.  Would be interesting to see a comparison of HAMMER+swapcache and ZFS+L2ARC, though.

-- Freddie Cashfjwcash at gmail.com
-- Freddie Cashfjwcash at gmail.com