According to Vladimir Litovka:
>
> I have a problem: for example, I access to any WWW-page and long time
> load it (about 5 minutes). After this time I think - 'Well, this page
> already in cache... he-he' and say to Explorer - 'Reload it!' And 8-O -
> squid begin load it directly from site!!! When I look into store.log, I
> see, that this URL was not put into cache:
>
> 838422634.292 RELEASE 200 0 [ 0s]
> 6995/6995 4/GET/http://www.netimg.com/TheHangar/Index.html
>
> There is really strange page, but I can't understand, in which cases squid
> will load URL from cache or directly from host?
Squid acts correctly with this page, because the pages from
www.netimg.com (at the one you mentioned) are not cacheable.
The page contains a Last-Modified: header, but the webserver is configured
to return the CURRENT date/time of the server as the LM-value
(Date: and Last-Modified: header are the same).
This makes the pages uncacheable, because squid calculates the value "0"
as the TTL and releases the page (IMHO).
I REALLY hate webmasters who hack there server to act like this. :-(
Maybe it would be a good idea to modify squid to check for such a
case and use the default TTL instead of the calculated TTL...
Regards...
-- Lars Eilebrecht sfx@www-cache.uni-siegen.deReceived on Sun Jul 28 1996 - 13:49:47 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:32:43 MST