[Gluster-users] two nodes with different performance metrics
John Madden
jmadden at ivytech.edu
Mon Jan 4 16:07:42 UTC 2010
Two-node AFR with two clients, all four machines identical, v2.0.9, etc.
I notice that one client is able to read and write much more quickly
than the other and I'm wondering how I go about finding why that is. In
a two-node AFR, is one node always elected to be the primary node for
writes and does that apply for reads too? That would explain my
situation (client0 and server0 share a hypervisor, client1 and server1
share a different hypervisor). If this behavior exists, can it be
configured?
bonnie++ output from the slower client:
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
pxhorde1 300M 23635 24 41146 3 1889 0 19153 18 30057 1
567.9 1
------Sequential Create------ --------Random
Create--------
-Create-- --Read--- -Delete-- -Create-- --Read---
-Delete--
files /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
/sec %CP
16 293 0 1807 0 346 0 270 0 1965 0
459 0
pxhorde1,300M,23635,24,41146,3,1889,0,19153,18,30057,1,567.9,1,16,293,0,1807,0,346,0,270,0,1965,0,459,0
and from the faster client:
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
pxhorde0 300M 27770 31 63545 6 4150 0 31808 29 66540 4
582.4 0
------Sequential Create------ --------Random
Create--------
-Create-- --Read--- -Delete-- -Create-- --Read---
-Delete--
files /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
/sec %CP
16 303 0 1714 0 253 0 295 0 1780 0
415 0
pxhorde0,300M,27770,31,63545,6,4150,0,31808,29,66540,4,582.4,0,16,303,0,1714,0,253,0,295,0,1780,0,415,0
--
John Madden
Sr UNIX Systems Engineer
Ivy Tech Community College of Indiana
jmadden at ivytech.edu
More information about the Gluster-users
mailing list