alit alit wrote: > > Dear sir/madam > > I am a software programmer and website administrator working in one of the > biggest websites in China. The page view of my company's website is top 10 > in china and top 150 in the world. Therefore, I want to use the squid as a > reverse proxy server to reduce the load of the webservers in my company's > website. > > But when I use the squid , I found that the squid is very slow !! It is > slower even then using apache directly. So I want to know if the squid just > can use in the small website? And it can't support more then 1500 > connections at the same time ? Or maybe I didn't do a corrent > configuration. > > this is my squid.conf and squid installed in red hat linux 6.2: > #visible name > visible_hostname 61.144.56.59 > memory_replacement_policy heap LFUDA > cache_replacement_policy heap LFUDA > # cache config: space use 10G and memory use 2048M > cache_dir ufs /home/squid/var/cache 9216 32 256 > cache_mem 700 MB > cache_effective_user nobody > cache_effective_group nobody > > http_port 80 > httpd_accel_host 61.140.254.153 > httpd_accel_single_host on > httpd_accel_port 80 > #httpd_accel_uses_host_header on > #httpd_accel_with_proxy on > > # accelerater my domain only > acl acceleratedHostA dstdomain .gznet.com > acl acceleratedHostB dstdomain .avl.com.cn > # accelerater http protocol on port 80 > acl acceleratedProtocol protocol HTTP > acl acceleratedPort port 80 > # access arc > acl all src 0.0.0.0/0.0.0.0 > no_cache allow all > # Allow requests when they are to the accelerated machine AND to the > # right port with right protocol > > # logging > emulate_httpd_log on > cache_store_log none > > # manager > http_access allow all > On which box/os/version ? How much memory ?
Your cache_mem is very (too?) high in value;try with 32M M.
