Le jeudi 10 septembre 2009 06:22:34, Amos Jeffries a écrit : > Muhammad Sharfuddin wrote: > > squid cache 2.7 STABLE5 > > squid cahce 3.0 STABLE19 > > > > I am using the following memory/cache replacement policy > > > > memory_replacement_policy heap LFUDA > > cache_replacement_policy heap LFUDA > > > > are they(heap LFUDA) best or most recommended ? or should I left them > > default. > > The defaults there were last tuned years ago and is not heap at all. > Read the descriptions of each carefully, the ones I saw a while back > covered several scenarios of traffic patterns and sizes and how the > policies worked in different network needs. > Pick the one which matches your visitor behaviour. Nobody else can > identify the best one for you with any accuracy. > > Amos > helo I did change my memory_replacement_policy, and mean_object_size because server is in production, a -k reconfigure is enoght or I must stop and start squid TIA LD
