dhottinger@xxxxxxxxxxxxxxxxxxxxxx wrote:
Quoting Adrian Chadd <adrian@xxxxxxxxxxxxxxx>:
On Tue, May 01, 2007, dhottinger@xxxxxxxxxxxxxxxxxxxxxx wrote:
I suddenly (last friday) started having issues when access google.com.
My access.log file shows all tcp_miss for google. Is anyone else
experiencing slow google access? I did get an email from google that
they were updating their applications (we use google calendars).
I've not heard about it. Do you have mime logging turned on so we can
see the headers w/ the request/reply?
Adrian
1178030214.203 538 10.40.15.123 TCP_MISS/200 5959 GET
http://tbn0.google.com/images? - DIRECT/72.14.211.104 image/jpeg ALL
OW "Visual Search Engine, Search Engines" [Accept:
*/*\r\nAccept-Language: en\r\nAccept-Encoding: gzip,
deflate\r\nCookie: PR
EF=ID=2377a880502b45e8:TM=1158762666:LM=1158762666:S=4J3pIHAN5lUxv6nf\r\nReferer:
http://images.google.com/images?q=newspaper
+comics+political+cartoons+on+the+war+in+Iraq&gbv=2&svnum=10&hl=en&start=40&sa=N&ndsp=20\r\nUser-Agent:
Mozilla/5.0 (Macintos
h; U; Intel Mac OS X; en) AppleWebKit/418.9 (KHTML, like Gecko)
Safari/419.3\r\nConnection: keep-alive\r\nHost: tbn0.google.c
om\r\n] [HTTP/1.0 200 OK\r\nContent-Type: image/jpeg\r\nServer:
btfe\r\nContent-Length: 5775\r\nDate: Tue, 01 May 2007 14:36:
53 GMT\r\nConnection: Keep-Alive\r\n\r]
Assuming you have not removed the default cache/no_cache line from your
squid.conf, anything with a question mark in the URL will not be cached.
From 2.6STABLE12's squid.conf.default:
#We recommend you to use the following two lines.
acl QUERY urlpath_regex cgi-bin \?
cache deny QUERY
2.5 is the same, except the directive is called no_cache.
In addition, the objects have no freshness information (Expires, or
Content-Length), so even without the explicit requirement within Squid
to not cache GET queries the listed objects are cacheable.
Chris
