Hi All, I set up some squid proxy as reverse proxy for serving large files (~50MB). If there are about 1200 concurrent connections (each connections persists about 2 mins, since I have file size ~50MB), The performance degrades quickly, and cache.log shows: 2008/04/14 22:48:11| comm_old_accept: FD 11: (53) Software caused connection abort 2008/04/14 22:48:11| httpAccept: FD 11: accept failure: (53) Software caused connection abort 2008/04/14 22:48:13| comm_old_accept: FD 11: (53) Software caused connection abort 2008/04/14 22:48:13| httpAccept: FD 11: accept failure: (53) Software caused connection abort 2008/04/14 22:48:14| comm_old_accept: FD 11: (53) Software caused connection abort 2008/04/14 22:48:14| httpAccept: FD 11: accept failure: (53) Software caused connection abort 2008/04/14 22:48:16| comm_old_accept: FD 11: (53) Software caused connection abort 2008/04/14 22:48:17| httpAccept: FD 11: accept failure: (53) Software caused connection abort I use squid-3.0.4 with kqueue on FreeBSD 7.0 RELEASE, with 4 CPUs and 4G RAM. Thank you for any suggestions to make the numbers better. My squid.conf: acl all src 0.0.0.0/0 acl localnet src 10.0.0.0/8 # RFC1918 possible internal network acl localnet src 172.16.0.0/12 # RFC1918 possible internal network acl localnet src 192.168.0.0/16 # RFC1918 possible internal network # acl SSL_ports port 443 acl Safe_ports port 80 # http acl Safe_ports port 21 # ftp acl Safe_ports port 443 # https acl Safe_ports port 70 # gopher acl Safe_ports port 210 # wais acl Safe_ports port 1025-65535 # unregistered ports acl Safe_ports port 280 # http-mgmt acl Safe_ports port 488 # gss-http acl Safe_ports port 591 # filemaker acl Safe_ports port 777 # multiling http acl CONNECT method CONNECT acl PURGE method PURGE acl manager proto cache_object acl localhost src 127.0.0.1/255.255.255.255 acl pixnet dstdomain .pixnet.net # TAG: http_access # INSERT YOUR OWN RULE(S) HERE TO ALLOW ACCESS FROM YOUR CLIENTS http_access allow manager localhost http_access deny manager http_access allow PURGE localhost http_access deny PURGE http_access allow pixnet http_access deny all icp_access allow localnet icp_access deny all #htcp_access allow localnet #htcp_access deny all # TAG: http_port http_port 80 accel defaultsite=server.pixnet.net # TAG: cache_peer # # proxy icp # # hostname type port port options # # -------------------- -------- ----- ----- ----------- # cache_peer parent.foo.net parent 3128 3130 proxy-only default # cache_peer sib1.foo.net sibling 3128 3130 proxy-only # cache_peer sib2.foo.net sibling 3128 3130 proxy-only cache_peer server.pixnet.net parent 80 0 no-query originserver # TAG: cache_mem (bytes) cache_mem 1024 MB cache_swap_low 80 cache_swap_high 95 maximum_object_size_in_memory 65536 KB # TAG: cache_dir cache_dir aufs /ad6/cache 102400 32 256 cache_dir aufs /ad4/cache 102400 32 256 minimum_object_size 0 KB maximum_object_size 131072 KB access_log /var/log/squid/access.log squid cache_log /var/log/squid/cache.log cache_store_log /var/log/squid/store.log logfile_rotate 30 # emulate_httpd_log off # pid_filename /usr/local/squid/logs/squid.pid # via on #accept_filter httpready memory_pools_limit 3608 MB forwarded_for on coredump_dir /ad6/cache cache_effective_user squid refresh_pattern . 1440 90% 2880 override-expire ignore-no-cache ignore-no-store minimum_expiry_time 2592000 seconds client_db off buffered_logs on half_closed_clients off #debug_options ALL,5 5,9