I'm using Squid as a transparent http cache for my dialup customer's.
It functions perfectly until it gets to the high water mark, at which
time it's fine until it's under load, 50+ request/sec, I'm using 2.2.5
Stable under linux 2.2.16 mostly, I've tried 2.3/2.4 series to no avail,
same problem with all of them. My hardware consist of Dual Celeron
466/512MB/4 20GB IBM Deskstar 7200RPM IDE w/promise controller card.
It's very strange, if the cache is under high water it works perfectly,
I've seen it do 70+ request/sec, but once it's at high water under load
it slows down response (it does recover after the peak period though),
is there anything I can do to fix this, or any advice on how to get it
to perform better. It's not swapping either, plenty of memory.
Regards
MIKE
-- To unsubscribe, see http://www.squid-cache.org/mailing-lists.htmlReceived on Tue Aug 22 2000 - 08:01:08 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:54:57 MST