Thank you! It seems there is an explanation for the behavior, which is good!

On 10/23/2015 12:44 AM, Gregory Farnum wrote:
On Thu, Oct 22, 2015 at 1:59 AM, Yan, Zheng <uker...@gmail.com> wrote:
direct IO only bypass kernel page cache. data still can be cached in
ceph-fuse. If I'm correct, the test repeatedly writes data to 8M
files. The cache make multiple write  assimilate into single OSD
write
Ugh, of course. I don't see a tracker ticket for that, so I made one:
http://tracker.ceph.com/issues/13569
-Greg

--
Goncalo Borges
Research Computing
ARC Centre of Excellence for Particle Physics at the Terascale
School of Physics A28 | University of Sydney, NSW  2006
T: +61 2 93511937

_______________________________________________
ceph-users mailing list
ceph-users@lists.ceph.com
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com

Reply via email to