Hi!
i checked the io performance with bonnie++ on an domU.
With "file:" i get a goot result (nearly the dom0 results):
Version 1.03 ------Sequential Output------ --Sequential Input-
--Random-
-Per Chr- --Block-- -Rewrite- -Per Chr- --Block--
--Seeks--
Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP
/sec %CP
server123 440M 52335 93 46364 3 214480 0 54105 77 709159 0
+++++ +++
------Sequential Create------ --------Random
Create--------
-Create-- --Read--- -Delete-- -Create-- --Read---
-Delete--
files /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
/sec %CP
16 20533 36 +++++ +++ 19130 10 23464 49 +++++ +++
16243 14
server123,440M,52335,93,46364,3,214480,0,54105,77,709159,0,+++++,+++,16,20533,36,+++++,+++,19130,10,23464,49,+++++,+++,16243,14
Then i try the same with just changed to "tap:aio:" and i get a much
slower result:
Version 1.03 ------Sequential Output------ --Sequential Input-
--Random-
-Per Chr- --Block-- -Rewrite- -Per Chr- --Block--
--Seeks--
Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP
/sec %CP
server123 440M 2125 3 2202 0 1907 0 28110 28 21798 0
13024 0
------Sequential Create------ --------Random
Create--------
-Create-- --Read--- -Delete-- -Create-- --Read---
-Delete--
files /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
/sec %CP
16 197 0 +++++ +++ 179 0 200 0 +++++ +++
160 0
server123,440M,2125,3,2202,0,1907,0,28110,28,21798,0,13023.9,0,16,197,0,+++++,+++,179,0,200,0,+++++,+++,160,0
_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users
On Thu, Jul 31, 2008 at 8:17 AM, MS <ms@toefke.com> wrote:> Hi! > > i checked the io performance with bonnie++ on an domU. > With "file:" i get a goot result (nearly the dom0 results):that "440M" is the test size? if so, it sounds too small. it should at least be twice (5-10x is better) the size of cache available. remember, "file:" devices might be cached by the Dom0, "tap:" won''t. with small files it''s common to see this pattern on most tests, but won''t benefit most real world uses. -- Javier _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Tastsite 440M needs over a hour with tap:aio: If i use some more i have to wait .... to christmas :) With file: the 440M need less then 3 minutes. I know, 440M is not a good value, but it never can be less then 1% of the performance before changeing. Javier Guerra schrieb:> On Thu, Jul 31, 2008 at 8:17 AM, MS <ms@toefke.com> wrote: > >> Hi! >> >> i checked the io performance with bonnie++ on an domU. >> With "file:" i get a goot result (nearly the dom0 results): >> > > that "440M" is the test size? if so, it sounds too small. it should > at least be twice (5-10x is better) the size of cache available. > remember, "file:" devices might be cached by the Dom0, "tap:" won''t. > with small files it''s common to see this pattern on most tests, but > won''t benefit most real world uses. > > >_______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users