Re: Performance question

On Wed, 16 Jun 1999, Scott Hess wrote:
]On Tue, 15 Jun 1999, Simon Rainey <srainey@rmplc.net> wrote:
]>[...]
]>I think everyone realises that Squid's achilles heal is the fact that it
]>relies on the undelying OS to manage disk files. While the tests I've
]>done are admittedly not very scientific, they do suggest that effort
]>spent on the rumoured Squid-FS (or similar) would be worthwhile.
]
]Hmm. You mentioned that you're running with 18 servers (presumably
]siblings)? If the FS is truly the limiting factor, it may be that running
]them all with proxy-only would help things out. The basic notion being
]that it would probably be faster to get something out of a sibling's
]in-memory cache than to get it off your disk.

During my peak times, I use the following config to help my toplevel
caches back onto their feet (as suggested by one of the Squid gurus):

        acl noonish time MTWHF 11:50-12:10
        acl noonish time MTWHF 12:50-13:10
        acl noonish time MTWHF 13:50-14:10
        acl noonish time MTWHF 14:50-15:10
        no_cache deny noonish

Le deagh dhùrachd,
Dipl.-Ing. Jens-S. Vöckler (voeckler@rvs.uni-hannover.de)
Institute for Computer Networks and Distributed Systems
University of Hanover, Germany; +49 511 762 4726
Received on Thu Jun 17 1999 - 04:59:52 MDT