博主不是专业的运维人员， 有时候新开业务服务器到生产环境忘记配置ubuntu的配置时候，就发生各种奇怪的悲剧比如：
redis出现大量TIMEWAIT状态，并且服务器不会收，原因是tcp_tw_recycle配置为0等等。为避免遗忘记录下新开的服务器必要的操作：
ubuntu网络优化
vi /etc/sysctl.conf
net.ipv4.tcp_max_tw_buckets = 5000
net.ipv4.tcp_syncookies = 1
net.ipv4.tcp_synack_retries = 2
net.ipv6.conf.all.disable_ipv6 = 1
net.ipv6.conf.default.disable_ipv6 = 1
net.ipv6.conf.lo.disable_ipv6 = 1
net.ipv4.tcp_orphan_retries = 3
net.ipv4.tcp_syncookies = 1
net.ipv4.tcp_tw_reuse = 1
net.ipv4.tcp_tw_recycle = 1
net.ipv4.tcp_fin_timeout = 3
net.ipv4.tcp_window_scaling = 1
net.ipv4.tcp_sack = 1
net.ipv4.tcp_keepalive_time = 30
net.ipv4.ip_local_port_range = 2000 65000
net.ipv4.tcp_max_tw_buckets
