Re: Questions

Yee Man Chan wrote:

> 1. In Aug 1 sv log, I have these two "busy" URLs.
>
> 164655179 http://sv.cache.nlanr.net:3128/squid-internal-periodic/
> store_digest
> 152234565 http://sv.cache.nlanr.net:3128/squid-internal-dynamic/
> netdb

These are internal between two Squids.

> 2. I found that there is some ftp URLs look like:
>
> ftp://ftp.dartmouth.edu/(null)
>
> What is the meaning of (null) at the end?

Most likely that something in the Squid FTP listing module was broken.
If I am not mistaken I sent a broken FTP patch to Duane somewhere around
this time (August) that could cause this effect..

If requests like this is found in logs from any Squid 2 releases then I
would be very interested in investigating those entries further.

> 1. Suppose a client stop an HTTP req at the middle of transfer. I know
> then the connection between Squid and that client would be closed. How
> about the connection between Squid and the web server? Is Squid always
> cache complete web objects? Would it cache incomplete object due to
> network failure between Squid and the web server?

See quick_abort*. By default it continues the download if it is a
cachable object.

> 2. Suppose there is a web object that is bigger than the maximum cache
> size. What does Squid do?

No idea what Squid 2 does. Duane?

> 3. For uncachable objects like URLs containing '?', Squid would
> not save them on the disk and always give users TCP_MISS/200. Is
> that true?

I think Squid 2 always stores them on disk.. Duane?

Squid 2 uses a different storage management than Squid 1.1.X. It is
supposed to be more like Squid 1.NOVM.. But I haven't really got the
time needed to investigate how Squid 2 really does it.

/Henrik
Received on Tue Jul 29 2003 - 13:15:54 MDT