Re: [squid-users] access.log

[...]
 
> > 2. Why does the amount of data delivered to the clients differ from one
> > response to another although
> > a) the same URL is accessed and
> > b) the respective web object is not refreshed ?
>
> It can be any of the following
>
> - Client sending an IMS query resulting in a 304 response
>
> - Client sending a range request, only asking for a piece of the object
>
> - Client aborts the request before the whole response have been sent
>
> > Is squid modifying any HTTP headers? Which ones?
>
> Squid adds to the Vary header, and strips hop-by-hop headers as required
> by the HTTP protocol. In addition you have optional anonymization
> functions which strips unwanted http headers.

According to my analysis this would lead to an abort rate of approx. 8-12%.
The differences in the amount of data transferred is <10 bytes in almost all
cases. Is this realistic?
 
Other possibilities include:
a) Squid stores multiple representations of an entity (-> server driven
content negotiation).
b) Squid modifies other end-to-end headers than the Vary header.
c) Hop-by-hop headers are also counted in the size field of access.log

Some comments would be fine!

Thanks in advance!
Regards,
 Michael Scheibel

-- 
Psssst! Mit GMX Handyrechnung senken: http://www.gmx.net/de/go/mail
100 FreeSMS/Monat (GMX TopMail), 50 (GMX ProMail), 10 (GMX FreeMail)