I wonder if anyone has a good I/O test which will sort of represent
the way squid needed the I/O to perform. Basically I need one program
which I can use to check the influence of various components of the
system ( OS, parameters, harddisk, library version ) on the I/O for
maximizing
squid performance.
For example, when I run on the ***SAME*** machine, one with a Fedora
Core 5 ( kernel 2.6.21.3 ) and one with uclibc-0.9.28 ( kernel 2.6.18 ),
using bonnie-1.4 ( a very old program and maybe buggy with respect to
the big size files ), this is the result :-
FC5 :
# ./Bonnie -s 2004 -d tmp
Bonnie 1.4: File 'tmp/Bonnie.24172', size: 2101346304, volumes: 1
Writing with putc()... done: 35399 kB/s 93.7 %CPU
Rewriting... done: 12557 kB/s 5.0 %CPU
Writing intelligently... done: 37587 kB/s 14.4 %CPU
Reading with getc()... done: 14878 kB/s 86.6 %CPU
Reading intelligently... done: 27846 kB/s -0.0 %CPU
Seeker 1...Seeker 2...Seeker 3...start 'em...done...done...done...
---Sequential Output (nosync)--- ---Sequential Input-- --Rnd
Seek-
-Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --04k
(03)-
Machine MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec
%CPU
testin 1*2004 35399 93.7 37587 14.4 12557 5.0 14878 86.6 27846 -0.0
130.3 -0.0
uclibc 0.9.28
# ./Bonnie -s 2004 -d tmp
Bonnie 1.4: File 'tmp/Bonnie.19932', size: 2101346304, volumes: 1
Writing with putc()... done: 44332 kB/s 35.9 %CPU
Rewriting... done: 18676 kB/s 11.3 %CPU
Writing intelligently... done: 49168 kB/s 18.5 %CPU
Reading with getc()... done: 39758 kB/s 30.5 %CPU
Reading intelligently... done: 44955 kB/s 13.8 %CPU
Seeker 1...Seeker 2...Seeker 3...start 'em...done...done...done...
---Sequential Output (nosync)--- ---Sequential Input-- --Rnd
Seek-
-Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --04k
(03)-
Machine MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec
%CPU
cnxsq 1*2004 44332 35.9 49168 18.5 18676 11.3 39758 30.5 44955 13.8 116.6
0.7
It seems that the I/O on uclibc-based system runs better than a FC5 system.
This is will opposite to my expectation. I am not sure if the conclusion
will
be applicable to running squid.
However, when I run the bonnie++-1.93c ( statically compiled version since
the
said uclibc system does not have C++ runtime ), the performance of both
configurations are almost identical.
Appreciate comments.
--------------------------------------------------------
Important Warning!
***************************
This electronic communication (including any attached files) may contain confidential and/or legally privileged information and is only intended for the use of the person to whom it is addressed. If you are not the intended recipient, you do not have permission to read, use, disseminate, distribute, copy or retain any part of this communication or its attachments in any form. If this e-mail was sent to you by mistake, please take the time to notify the sender so that they can identify the problem and avoid any more mistakes in sending e-mail to you. The unauthorised use of information contained in this communication or its attachments may result in legal action against any person who uses it.
