Reuben Farrelly wrote:
> Could you also explain the difference between the minimum and maximum values
> for the refresh time?
They are just that, A minimum and maximum value for the percentage
calculation.
So lets assume the refresh pattern of 5 20% 1440
This says that objects should be considered fresh for 20% of their
modification age, but no less than 5 minutes and no more than 24 hours
(1440 minutes).
However, you should NEVER use a min age other than 0 unless you know
exacly what you are doing. A lot of dynamic content sites not designed
to operate with caching will behave strangely of you force them to be
cached by a min age, and it also goes against the recommendations of
HTTP/1.1.
> When squid is making a decision about the freshness of an object, how does
> it use the two values? Would not one value be enough to determine freshness
> since an object is either "fresh" or "stale"?
As I said above, the two values are the outer limits of the age based
freshness calculation. If you are only interested in the age then set
min to 0 and max to a huge number. If you want everything cachable to be
cached for 60 minutes then set both to 60 (the percentage then won't
matter, as the freshness range is collapsed into one point where
everything is fresh for 60 minutes)
Normally you are only interested in tuning the upper limit, to guarantee
that objects matching a certain pattern are never more than X minutes
stale irregardless of their last modification-time when the object was
cached.
-- Henrik Nordstrom Squid hackerReceived on Mon Feb 07 2000 - 16:56:55 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:51:01 MST