Mailing-List: contact users-help@httpd.apache.org; run by ezmlm
Precedence: bulk
Reply-To: users@httpd.apache.org
Received-SPF: pass (nike.apache.org: local policy)
DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws;
  s=s1024; d=yahoo.com;
  h=Message-ID:X-YMail-OSG:Received:X-Mailer:References:Date:From:Subject:To:In-Reply-To:MIME-Version:Content-Type:Content-Transfer-Encoding;
  b=IFJKXplEVLX8hACqNJ0joV3Du3woy1B2WMoTwnSpTCyFt0hSSdccXA5EgX2xJ71jmc/ZbktaunyBXFlgZYPBFLwHIoYw1pbEElAtds45eoCdKjeG49rlKxYrVLVRYJrD7Qa8k8rWTkx/OY9lBNYwiU1giDatFM4ZkuuOm5GuUWk=;
Message-ID: <41438.76090.qm@web113609.mail.gq1.yahoo.com>
References: <737795.26418.qm@web112809.mail.gq1.yahoo.com>
 <4D947D3E.4020903@catseye.org>
Date: Thu, 31 Mar 2011 08:16:45 -0700 (PDT)
From: Tony Anecito <adanecito@yahoo.com>
To: users@httpd.apache.org
In-Reply-To: <4D947D3E.4020903@catseye.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: quoted-printable
Subject: Re: [users@httpd] What is %D in access log meassuring?

Hi All,=0A=0AI am confused by these messages because earlier %D was defined=
 to include the =0Atime for the processing all lines after the first one of=
 the packets containing =0Athe request and the time to process the request =
itself. If true then the =0Afollowing emails about checking the network and=
 ipB itself make sense.=0A=0AAlso, any request will takeup CPU resources on=
 the server and other servers =0Adepending upon where the image is hosted a=
nd if is cached. The only case where =0Amaybe the request does not takeup d=
irect server CPU time is if the request goes =0Athrough an intellegent netw=
ork card and it is smart enough to know to serve the =0Aimage file from it'=
s cache but I am not sure those cards can do that.=0A=0AI have seen Apache =
Components group performance test results that indicate the =0Aresponse tim=
e for a single request where there is no other requests occuring is =0Aunde=
r a millisecond and that includes going through a servlet. I have confirmed=
 =0Amyself for a servlet containing no code using the Apache APR with Tomca=
t 7.0.8 =0Athe response as measured on at the client running on the server =
Tomcat is =0Alocated using uri localhost is under 600 microseconds. To my u=
nderstanding that =0Abypasses the network card. If I ran the same test from=
 a cheap laptop running on =0Amy 1Gb Ethernet I get around 2-3ms.=0A=0ARega=
rds,=0A-Tony=0A=0A=0A=0A----- Original Message ----=0AFrom: Mark Montague <=
mark@catseye.org>=0ATo: users@httpd.apache.org=0ACc: Janne H <jannehson51@y=
ahoo.com>=0ASent: Thu, March 31, 2011 7:10:22 AM=0ASubject: Re: [users@http=
d] What is %D in access log meassuring?=0A=0AOn March 31, 2011 8:37 , Janne=
 H <jannehson51@yahoo.com>=A0 wrote:=0A> Well, I'm still a little confused.=
=0A> I'm trying to find out why the accesslog shows the line=0A> =0A> =0A> =
ipA [31/Mar/2011] "GET /file.jpg HTTP/1.0" 200 42981 "-" "ApacheBench/2.3" =
=0A3560=0A> ipB [31/Mar/2011] "GET /file.jpg HTTP/1.0" 200 42981 "-" "Apach=
eBench/2.3" =0A>93574=0A> =0A> (the ipA is much "closer" to the server than=
 ipB).=0A=0AFirst, are these results consistently repeatable over a large n=
umber of =0Ameasurements?=A0 If not, then it is likely just normal statisti=
cal variation due =0Ato normal network traffic, CPU scheduling, or somethin=
g similar.=A0 Otherwise...=0A=0AWhat does 'much "closer"' mean?=0A=0AWhat i=
s the network topology (all links, hardware, configuration) between ipA =0A=
and your web server?=A0 Between ipB and your web server?=0A=0AWhat is the h=
ardware (number and speed of CPUs, amount of RAM, etc.) for ipA and =0AipB?=
=A0 What OS and version are each running?=A0 How heavily loaded is each one=
 =0Aduring the test (CPU, I/O)?=A0 How is the kernel and networking stack o=
n each one =0Aconfigured?=0A=0AI recommend using a network sniffer to analy=
ze differences in network traffic =0Afor a request sent from ipA versus a r=
equest sent from ipB.=A0 Try and find out =0Awhere the delay is occurring f=
or ipB.=0A=0AIf (and only if) the network trace shows that the delay is in =
waiting for your =0Aserver to generate and send packets to ipB -- as oppose=
d to your server waiting =0Afor packets from ipB -- then use performance an=
alysis tools on your server such =0Aas strace, truss, DTrace, or SystemTap =
to figure out what your kernel, and web =0Aserver processes are doing durin=
g that time.=0A=0AKeep in mind that you're looking at a 90 millisecond (90,=
000 microsecond) =0Adifference between the two clients.=A0 Depending on you=
r situation, this may or =0Amay not actually be a significant real-world pr=
oblem, and hence it may or may =0Anot justify a large amount of time and ef=
fort for finding the cause.=0A=0A=0A> i also tried a curl download with a l=
imit on 1000 bytes/s but the time logged =0A>in access-log many orders of m=
agnitute away from the acctual time it took to =0A>receive the image (about=
 42 seconds) So I guess the time that is logged is "when =0A>apache has sen=
t the final bytes to the level below in the network stack?=0A=0A%D measures=
 how long Apache spends processing the request.=A0 It does not measure =0Ah=
ow long other things on your server (kernel, host-based firewall, etc.) tak=
e to =0Aprocess the data, nor does it measure how long it takes data to act=
ually reach =0Athe client.=A0 If you are interested in how long the request=
 takes on the client =0A(including how long it takes the response to travel=
 from the server to the =0Aclient), then you will need to measure that on t=
he client, not on the server.=0A=0A--=0A=A0 Mark Montague=0A=A0 mark@catsey=
e.org=0A=0A=0A-------------------------------------------------------------=
--------=0AThe official User-To-User support forum of the Apache HTTP Serve=
r Project.=0ASee <URL:http://httpd.apache.org/userslist.html> for more info=
.=0ATo unsubscribe, e-mail: users-unsubscribe@httpd.apache.org=0A=A0 "=A0 f=
rom the digest: users-digest-unsubscribe@httpd.apache.org=0AFor additional =
commands, e-mail: users-help@httpd.apache.org

---------------------------------------------------------------------
The official User-To-User support forum of the Apache HTTP Server Project.
See <URL:http://httpd.apache.org/userslist.html> for more info.
To unsubscribe, e-mail: users-unsubscribe@httpd.apache.org
   "   from the digest: users-digest-unsubscribe@httpd.apache.org
For additional commands, e-mail: users-help@httpd.apache.org