> -----Original Message-----
> A disturbing number of these downloads are "FlashGet" and "DA"
> type apps that open absurd numbers (as many as 32) simultaneous
> byterange requests. I am thinking this has something to do with
> all those Range: requests interacting badly with the cache.
As an update to this:
I forced we_do_ranges to 0 in http.c and this did not alter the bad
behavior. Only no_cache seems to defeat it, and that is not a good solution
for a cache!: )
Perhaps the stalls aren't so much related to Range requests as to the number
of people requesting the same cache object? How can this theory be
objectively proved/disproved?
Thanks,
Jeff
-- NearlyFreeSpeech.NET Support Team support@nearlyfreespeech.net http://www.nearlyfreespeech.netReceived on Sat Mar 16 2002 - 14:18:21 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:06:57 MST