i'm not that fan of linking to web pages so you will need to read but
sometimes it's a must.
starts with this:
http://www.mnot.net/cache_docs/
to understand the basic things about http cache.
this page:
http://www.squid-cache.org/Doc/config/refresh_pattern/
is specific on squid refresh patterns usage.
i must say that the above squid document took me awhile to understand
and until today i dont like the way it's organized (just my opinion) but
it contains everything about refresh_pattern.
this document:
http://etutorials.org/Server+Administration/Squid.+The+definitive+guide/Chapter+7.+Disk+Cache+Basics/7.7+refresh_pattern/
was taken from squid book and is organized nicely.
i will give you one site that is organized as heaven for http cache:
http://www.djmaza.com/
about the frontera.info you will want to check the main page instead of
the one you checked using:
http://redbot.org/?descend=True&uri=http://www.frontera.info/Home.aspx
the server time is not correct but as for cache age it's not important.
Regards,
Eliezer
(you can look me up on squid irc channel)
On 24/05/2012 03:59, Beto Moreno wrote:
Hi, thanks for your info.
I had try that tool, just need to understand:
frontera.info say:
General
The server's clock is 3 min 58 sec behind.
Content Negotiation
The resource doesn't send Vary consistently.
The server's clock is 3 min 58 sec behind.
Content negotiation for gzip compression is supported, saving 42%.
The server's clock is 3 min 58 sec behind.
Caching
This response only allows a private cache to store it.
This response allows a cache to assign its own freshness lifetime.
Now the embedded:
Problems
The server's clock is 3 min 58 sec behind.
This response allows a cache to assign its own freshness lifetime.
The resource doesn't send Vary consistently.
The Content-Disposition header doesn't have a 'filename' parameter.
Cache-Control: public is rarely necessary.
The If-Modified-Since response is missing required headers.
Now yahoo.com
General
The server's clock is correct.
Caching
This response only allows a private cache to store it.
This response allows a cache to assign its own freshness lifetime.
Now: noticiasmvs.com
General
The server's clock is correct.
The Content-Length header is correct.
Content Negotiation
Content negotiation for gzip compression is supported, saving 22%.
Caching
This response allows all caches to store it.
This response allows a cache to assign its own freshness lifetime.
What I understand is that yahoo/frontera won't let squid to save some
of their data, and noticiasmvs is open for squid, right?
Will very appreciated if someone could explain me a little more about
this output from this site I want to go deeper with squid, what we can
do in this situation(private cache)?
Thanks!!!
On Wed, May 23, 2012 at 2:57 AM, Eliezer Croitoru<eliezer@xxxxxxxxxxxx> wrote:
you can try to use this tool:
http://redbot.org/
to make sure what are the sites cachebilaty options
maybe some objects there need some cache enforcement rules in the
refresh_pattern specified for them.
Eliezer
On 23/05/2012 02:54, Beto Moreno wrote:
I had been working on the settings:
refresh_pattern.
The doc say that is better for the new websites that use dynamic
content and a friend here at the list explain me the difference.
My test was simple:
use 2 browsers: firefox/iexplore.
Run the test twice for each site.
first run
firefox site1, site2,site3,site4
iexplore site1, site2,site3,site4
run ccleaner, repeat the test.
run srg to get my squid-cache peformance and free-sa.
They where 3 settings I try and make the same test.
NOTE: every time I start a setting, I delete my cache, clean my logs
and start from 0.
setting 1 default settings
acl QUERY urlpath_regex cgi-bin \?
cache deny QUERY
setting 2 new way:
disable the old way:
#acl QUERY urlpath_regex cgi-bin \?
#cache deny QUERY
refresh_pattern -i (/cgi-bin/|\?) 0 0% 0
refresh_pattern . 0 20% 4320
setting 2:
refresh_pattern -i (/cgi-bin/|\?) 0 0% 0
refresh_pattern -i \.(gif|png|jpg|jpeg|ico)$ 10080 90% 43200
refresh_pattern -i \.index.(html|htm)$ 0 40% 10080
refresh_pattern -i \.(html|htm|css|js)$ 1440 40% 40320
refresh_pattern . 0 20% 4320
Them after I finish my test I start reviewing my logs and compare,
the sites I use was:
yahoo.com
osnews.com
frontera,info(local news paper)
noticias,nvs.com
centos.org
I didn't interact with the site, just get to the first page, finish
loading and done, continue with the next one.
Once I check my reports I didn't see to much difference, I found just
1 log that the old way didn't "cache" 1 thing, check:
setting 1/2 have this:
1337667655.898 0 192.168.50.100 TCP_MEM_HIT/200 21280 GET
http://www.frontera.info/WebResource.axd? - NONE/-
application/x-javascript
setting 1 TCP_MISS.
Example of part my logs:
1337667655.596 43 192.168.50.100 TCP_MISS/302 603 GET
http://frontera.info/ - DIRECT/216.240.181.163 text/html
1337667655.748 54 192.168.50.100 TCP_MISS/200 1454 GET
http://www.frontera.info/HojasEstilos/Horoscopos.css -
DIRECT/216.240.181.163 text/css
1337667655.749 52 192.168.50.100 TCP_MISS/200 1740 GET
http://www.frontera.info/Includes/Controles/LosEconomicos.css -
DIRECT/216.240.181.163 text/css
1337667655.749 49 192.168.50.100 TCP_MISS/200 1557 GET
http://www.frontera.info/Includes/Controles/ReporteroCiudadano.css -
DIRECT/216.240.181.163 text/css
1337667655.754 54 192.168.50.100 TCP_MISS/200 1697 GET
http://www.frontera.info/Includes/Controles/Elementos.css -
DIRECT/216.240.181.163 text/css
1337667655.780 24 192.168.50.100 TCP_MISS/200 1406 GET
http://www.frontera.info/Includes/Controles/Finanzas.css -
DIRECT/216.240.181.163 text/css
1337667655.817 124 192.168.50.100 TCP_MISS/200 21639 GET
http://www.frontera.info/HojasEstilos/Estilos2009.css -
DIRECT/216.240.181.163 text/css
1337667655.898 0 192.168.50.100 TCP_MEM_HIT/200 21280 GET
http://www.frontera.info/WebResource.axd? - NONE/-
application/x-javascript
1337667655.903 20 192.168.50.100 TCP_MISS/200 1356 GET
http://www.frontera.info/Interactivos/lib/jquery.jcarousel.css -
DIRECT/216.240.181.163 text/css
1337667655.907 308 192.168.50.100 TCP_MISS/200 116552 GET
http://www.frontera.info/Home.aspx - DIRECT/216.240.181.163 text/html
1337667655.935 23 192.168.50.100 TCP_MISS/200 3934 GET
http://www.frontera.info/Interactivos/skins/fotos/skin.css -
DIRECT/216.240.181.163 text/css
1337667655.966 27 192.168.50.100 TCP_MISS/200 3995 GET
http://www.frontera.info/Interactivos/skins/elementos/skin.css -
DIRECT/216.240.181.163 text/css
1337667655.971 23 192.168.50.100 TCP_MISS/200 4260 GET
http://www.frontera.info/HojasEstilos/ui.tabs.css -
DIRECT/216.240.181.163 text/css
1337667655.972 24 192.168.50.100 TCP_MISS/200 4953 GET
http://www.frontera.info/HojasEstilos/thickbox.css -
DIRECT/216.240.181.163 text/css
1337667655.993 21 192.168.50.100 TCP_MISS/200 4380 GET
http://www.frontera.info/js/finanzas.js - DIRECT/216.240.181.163
application/x-javascript
1337667655.997 47 192.168.50.100 TCP_MISS/200 9341 GET
http://www.frontera.info/Interactivos/lib/jquery.jcarousel.pack.js -
DIRECT/216.240.181.163 application/x-javascript
1337667656.023 25 192.168.50.100 TCP_MISS/200 4239 GET
http://www.frontera.info/videos/external_script.js -
DIRECT/216.240.181.163 application/x-javascript
3 settings same TCP_MISS.
I was thinking that maybe I will get more TCP_HIT, MEM_HIT, but no.
noticiasmvs.com a lot HIT's but with the 3 settings.
do this site disable caching their site? exist a way to find out?
what could cause to still get a lot of MISS?
where my settings wrong?
my test was not the best way?
how can I see if this new settings make a difference?
Any input will be appreciated, thanks for your time!!!
I'm using squid 2.7.x
--
Eliezer Croitoru
https://www1.ngtech.co.il
IT consulting for Nonprofit organizations
eliezer<at> ngtech.co.il
--
Eliezer Croitoru
https://www1.ngtech.co.il
IT consulting for Nonprofit organizations
eliezer <at> ngtech.co.il
