> Your squid.conf default settings might have "acl QUERY url_regex cgi-bin
> \?", "cache deny QUERY", or the nicer "refresh_pattern -i (/cgi-bin/|\?)
> 0 0% 0" to tell squid explicitly what requests to consider dynamic. But
> the defaults are not involving .shtml
I did not find any "cache deny QUERY" in squid.conf. I have the default refresh_pattern settings:
#Suggested default:
refresh_pattern ^ftp: 1440 20% 10080
refresh_pattern ^gopher: 1440 0% 1440
refresh_pattern (cgi-bin|\?) 0 0% 0
refresh_pattern . 0 20% 4320
> The website itself is being nasty and telling Squid that those whole
> pages change between browsers. "Vary: User-Agent" which can make Squid
> drop objects when a different browser from the one which caused storage
> earlier requests it.
> NP2: Internet Explorer is particularly bad with this since each PC users
> installed Ad-Ons, their versions, the patches applied to IE in that box
> and that users account, and the exact build time of the IE binary are
> all stored in the User-Agent text.
I do not see anything unusual in the page source. Here is that metadata:
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
I just did a test: loaded the page in Firefox, closed the tab, opened a new tab and loaded the page in Firefox again. squid reported a TCP_MISS on the shtml page and a TCP_IMS_HIT on a CSS file. It appears Firefox cached all the other objects on the page except for these two. Strangely, store.log shows a RELEASE for both files. I will do some more testing and check out any squid debug options.
Thanks, Norbert
