Douglas Bell wrote:
Alright, thanks. I was finally shown the right settings to increase the timeout
to 60 seconds, so this problem should hopefully disappear (or at least be quite
reduced).
----
Have those setting been implemented yet?
I got these on 2 different browsers this morning:
[0507_122816.00] 15426ms; ln=814 (-0/53) MISS/503 <Athenae [GET
http://www.cpantesters.org/distro/P/P.html?version=1.1.36 -
23.235.47.204 text/html]
+75.42 15359ms; ln=814 (11/53) MISS/503 <Athenae [GET
http://www.cpantesters.org/distro/P/P.html?version=1.1.36 -
23.235.47.204 text/html]
+0.57 512ms; ln=1577 (2.7K/3.0K) REFRESH_UNMODIFIED/200 <Athenae
[GET http://www.cpantesters.org/favicon.ico - 23.235.47.204
image/vnd.microsoft.icon]
The info at the beginning of the line in brackets is a date+time, while
a "+75.42" means that entry was logged 75.42 seconds later.
The 2nd number is the round-trip request time (~15 seconds in each
case).
Looking at my squid.conf file, I see a default setting of 1 minute
for a connect timeout, and a 150 second timeout for requests. Just
to answer the potential question of whether or not 150 could mean
tenths: unlikely. Units are specified in the config file as
part of the number, i.e:
# read_timeout 15 minutes
request_timeout 150 seconds
It looks like it might have checked to see if the favicon was modified
by asking for it's last modified time/date. I just started 'IE' to see
if it could fetch the data, and I _think_ it checks content once/session
even if a locally cached copy indicates it "should" be fresh.