Re: no_cache tag..

From: Henrik Nordstrom <hno@dont-contact.us>
Date: Tue, 08 Feb 2000 11:16:06 +0100

Reuben Farrelly wrote:

> Thanks for the explanation. The idea of the min and max age being
> "boundaries" centred around the % figure helps understand the concept. The
> squid.conf comments don't explain too well how the min and max are
> essentially "boundaries" that the percentage freshness value cannot exceed.
> I had the idea that they were independent values and not related to the
> percent (obviously I was wrong).

Now that you remind me I have to admit that I was partially wrong.

The min age is a lot trickier. The exact meaning of it depends on what
other options you have in the refresh_pattern.

The default meaning (without any options) is to act as a default
freshness TTL if the reply does not contain any other freshness related
information. This is also why it breaks dynamic applications.

Then there are two options which makes the min age even more aggressive:

override-expire causes the min age to override any expiry information
sent by the origin server. DON'T USE THIS UNLESS YOU KNOW WHAT YOU ARE
DOING!

override-lastmod causes the min age to be used as the lower limit for
the % based LMT calculation.

Squid should probably be changed to by default have the min age act as
only the lower limit for LMT calculations, and options for using it as a
default TTL or not override LMT.

/Henrik
Received on Tue Feb 08 2000 - 06:41:25 MST

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:51:01 MST