We are trying to set up the following scenario:
a root squid cache with two virtual interfaces: proxy and fastproxy
a number of seconday level caches which should be able to fetch from
'proxy' as a parent thru a delay-pool configuration and from 'fastproxy'
as a sibling with no delay-pools.
Our goal is to send cache hits at wire speed while shaping cache-misses
to some desiderable bandwidth.
We have thus implemented this idea following the indications given in
the squid.conf file and in the FAQ.
there is an ACL which prevents client caches to fetch misses from
fastproxy
acl fastproxy myip 212.aaa.bbb.ccc/255.255.255.255
miss_access deny fastproxy
miss_access allow all
The root proxy doesn't complain about the conf file...
client caches have two cache_peer lines..
one with a parent setting for proxy
one with a sibling setting for fastproxy, with higher priority.
We thought the client cache should check with the sibling cache at first
and then, in case of a miss, proceed to the slower parent cache...
in real word after a few requests the client caches disable the sibling
peer and proceed normally with the slow parent.
What's wrong... is the idea or the implementation ?
thankx,
Paolo Pennisi
-- Paolo Pennisi silos@zero.it http://www.zero.it ___ ZeRO Computing S.r.l. - Software & Silicon Solutions ___ via A. Cecchi, 8 20146 Milano ITALY - tel. +39-02-48029744 fax. +39-02-48029583Received on Thu Feb 10 2000 - 12:26:44 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:51:10 MST