Настройки для ядра >= 3.2 (Впрочем, абсолютно наверняка и для 2.6 подойдут) сервера рассчитанного на сетевую нагрузку (router/nat/web/etc). В данном случае это xeon E5420, 8G RAM.
# SWAP usage vm.swappiness = 60 # Controls the maximum size of a message, in bytes kernel.msgmnb = 65536 # Controls the default maxmimum size of a mesage queue kernel.msgmax = 65536 # Controls the maximum shared segment size, in bytes kernel.shmmax = 4194349056 # Controls the maximum number of shared memory segments, in pages kernel.shmall = 1024011 ################## ### NETWORKING ### ################## # Max sockets who wait for connections net.core.somaxconn = 16384 # Max queue packets net.core.netdev_max_backlog = 8192 # Disable ipv6 net.ipv6.conf.all.disable_ipv6 = 1 # Ignore ICMP broadcasts net.ipv4.icmp_echo_ignore_broadcasts = 1 # Cisco ARP problem solving net.ipv4.conf.lo.arp_ignore = 1 net.ipv4.conf.lo.arp_announce = 1 net.ipv4.conf.all.arp_ignore = 1 net.ipv4.conf.all.arp_announce = 1 # Tunneling and packet forwarding for IPv4 net.ipv4.conf.all.rp_filter = 0 net.ipv4.conf.default.rp_filter = 0 net.ipv4.ip_forward = 1 # Max oprhan sockets without handle net.ipv4.tcp_max_orphans = 65536 # Count fails before destroy TCP-connection, closed local net.ipv4.tcp_orphan_retries = 0 # Time through which sockets in state FIN-WAIT-2 will be closed net.ipv4.tcp_fin_timeout = 10 # Max connections without confirmation net.ipv4.tcp_max_syn_backlog = 4096 # How often check SO_KEEPALIVE state connections net.ipv4.tcp_keepalive_time = 1800 # Retries count transmit SYNACK packets before drop connection net.ipv4.tcp_synack_retries = 2 # Enable tcp_congestion_control net.ipv4.tcp_timestamps = 1 # Protocol used to control the load networks net.ipv4.tcp_congestion_control = htcp # What ports we can use net.ipv4.ip_local_port_range = 1024 65535 # TIME_WAIT sockets quick utilization net.ipv4.tcp_tw_recycle = 1 # We can reuse TIME_WAIT sockets net.ipv4.tcp_tw_reuse = 1 # (tcp_keepalive_intvl * tcp_keepalive_probes) before connection will be disconnected when there is no response net.ipv4.tcp_keepalive_probes = 3 net.ipv4.tcp_keepalive_intvl = 15 # We do not use SACK in out TCP stack net.ipv4.tcp_sack = 0 # And D-SACK too net.ipv4.tcp_dsack = 0 # # Before use conntrack, we must load this module in our core! # https://www.kernel.org/doc/Documentation/networking/nf_conntrack-sysctl.txt # # Default for generic timeout (Def 600) net.netfilter.nf_conntrack_generic_timeout = 30 # Time out for established connection (Def 432000 (5 days)) net.netfilter.nf_conntrack_tcp_timeout_established = 300 # Timeout for SYN packets (Def 120) net.netfilter.nf_conntrack_tcp_timeout_syn_sent = 30 net.netfilter.nf_conntrack_tcp_timeout_syn_recv = 30 # Timeout for TIME_WAIT connections (Def 120) net.netfilter.nf_conntrack_tcp_timeout_time_wait = 30 # Timeout for FIN_WAIT connections (Def 120) net.netfilter.nf_conntrack_tcp_timeout_fin_wait = 30 # Timeout for CLOSE_WAIT connections (Def 60) net.netfilter.nf_conntrack_tcp_timeout_close_wait = 15 # Max count our connections (sysctl net.netfilter.nf_conntrack_count) net.netfilter.nf_conntrack_expect_max = 4096 net.netfilter.nf_conntrack_max = 1048576
[…] ядро. Можно использовать мой strong optimization из предыдущей […]
WE ᎳILL.? They each shouted they usually ran to the
bed room bickerinng about who will get to go first.
wtf?
http://www.sncmedia.ru/upload/iblock/5c2/5c2f1214566e00eb095c1db6e446532f.jpg