> On Fri, 2 Apr 2004, Matus UHLAR - fantomas wrote:
> > I'd like to split my squid spool to 2 filesystems: one large filesystem
> > with large block sizes for large files, other small filesystem with small
> > block sizes for small files.
On 02.04 23:29, Henrik Nordstrom wrote:
> Sound reasonable. Would allow for a large cache for large downloaded
> objects without exploding in memory usage due to a huge number of small
> objects being cached..
> > 2. define min-size option for cache_dir
...
> > I've looked at sources and I think min-size would be easy to implement...
> no harder than max-size.
>
> for both there will be some objects which are uncertain (where the full
> object size is not known before the complete object has been downloaded).
So, we have the question here: how is this problem solved now?
Also, are there any info how often (% of requests/responses) do we not
know the size of object before fetching them?
(just for info, I think if we can't know the size, we will just ignore the
min-size option)
-- Matus UHLAR - fantomas, uhlar@fantomas.sk ; http://www.fantomas.sk/ Warning: I wish NOT to receive e-mail advertising to this address. Varovanie: na tuto adresu chcem NEDOSTAVAT akukolvek reklamnu postu. Fucking windows! Bring Bill Gates! (Southpark the movie)Received on Mon Apr 05 2004 - 03:59:11 MDT
This archive was generated by hypermail pre-2.1.9 : Fri Apr 30 2004 - 12:00:01 MDT