"Steve Campbell" <scampbel@iqtinc.com> writes:
> Has any one tried a very large refresh parameter list, with say 1000
> instances in it? I was interested in tailoring the rules to a large
> number of very specific objects, hopefully turning REFRESH_HITS into
> TCP_HITS without substantially increasing the risk of returning stale
> objects. Is there much processor or memory overhead in doing so?
I've never tried that but I'd be interested in seeing a two step approach to
"staling": if an object is "almost stale", than when requested, the cached
object will be returned immediately (as if it was fresh), but a refresh will be
initiated in the background (to make sure the served object was indeed fresh).
The "almost stale" state could be defined implicitely as whatever fraction
of your "refresh" setting. Or maybe a separate configuration directive
could control it.
Stefan
Received on Thu Oct 30 1997 - 11:27:47 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:37:23 MST