大量LAST_ACK 分析过程

来源:本站原创 Linux 超过1,354 views围观 0条评论

现象:在netstat的时候发现大量处于LAST_ACK状态的TCP连接,达到在ESTABLISHED状态的90%以上[root@ccsafe ~]# netstat -ant|fgrep ":"|cut -b 77-90|sort |uniq -c                 
       6 CLOSE_WAIT
       7 CLOSING      
    6838 ESTABLISHED
    1037 FIN_WAIT1   
     357 FIN_WAIT2   
    5830 LAST_ACK     
       2 LISTEN      
     276 SYN_RECV     
      71 TIME_WAIT   
[root@ccsafe ~]#
看看系统状态,性能都花在系统中断和上下文切换
[root@ccsafe ~]# vmstat 2
procs ———–memory———- —swap– —–io—- –system– —–cpu——
r b    swpd    free    buff cache    si    so     bi     bo    in    cs us sy id wa st
1 0       0 3091812 363032 284132     0     0      0      0     1     1 0 0 100 0 0
0 0       0 3091812 363032 284132     0     0      0      0 13750 3174 0 5 94 0 0
0 0       0 3091936 363032 284132     0     0      0      0 13666 3057 1 5 94 0 0
0 0       0 3092060 363032 284132     0     0      0     16 13749 3030 0 5 95 0 0
0 0       0 3092060 363032 284132     0     0      0      0 13822 3144 0 5 95 0 0
0 0       0 3092060 363032 284132     0     0      0      0 13390 2961 0 5 95 0 0
0 0       0 3092060 363032 284132     0     0      0      0 13541 3182 0 6 94 0 0
查看socket队列信息
[root@ccsafe ~]# sar -n SOCK 5
Linux 2.6.18-53.1.13.el5PAE (ccsafe)      10/21/2008
06:31:43 PM     totsck     tcpsck     udpsck     rawsck    ip-frag     tcp-tw
06:31:48 PM       6951      13868          1          0          0        430
Average:          6951      13868          1          0          0        430
根据TCP状态的变化过程来分析,LAST_ACK属于被动关闭连接过程中的状态
ESTABLISHED->CLOSE_WAIT->(发送ACK)->LAST_ACK->(发送FIN+接收ACK)->CLOSED
现在状态都堆积到LAST_ACK,初步判断问题从上下两个状态着手
调节一下LAST_ACK时间…
[root@ccsafe ~]# sysctl -a |grep last_ack
net.ipv4.netfilter.ip_conntrack_tcp_timeout_last_ack = 30
[root@ccsafe ~]# sysctl -w net.ipv4.netfilter.ip_conntrack_tcp_timeout_last_ack=10
net.ipv4.netfilter.ip_conntrack_tcp_timeout_last_ack = 10
[root@ccsafe ~]# sysctl -p
[root@ccsafe ~]# watch -n 10 "netstat -ant|fgrep ":"|cut -b 77-90|sort |uniq -c"
Every 5.0s: netstat -ant|fgrep :|cut -b 77-90|sort |uniq -c                     
       6 CLOSE_WAIT
       9 CLOSING
    6420 ESTABLISHED
     693 FIN_WAIT1
     391 FIN_WAIT2
    5081 LAST_ACK
       2 LISTEN
     203 SYN_RECV
      66 TIME_WAIT
检查一下LAST_ACK所对应的应用
[root@ccsafe ~]# netstat -ant|fgrep "LAST_ACK"|cut -b 49-75|cut -d ":" -f1|sort |uniq -c|sort -nr –key=1,7|head -5
     101 220.160.210.6
      46 222.75.65.69
      31 221.0.91.118
      24 222.210.8.160
      22 60.161.81.28
[root@ccsafe ~]#
[root@ccsafe ~]# netstat -an|grep "220.160.210.6"
tcp         0 17280 10.1.1.145:80            220.160.210.6:52787          ESTABLISHED
tcp         1 14401 10.1.1.145:80            220.160.210.6:52513          LAST_ACK     
tcp         1 14401 10.1.1.145:80            220.160.210.6:52769          LAST_ACK     
tcp         1 14401 10.1.1.145:80            220.160.210.6:52768          LAST_ACK     
tcp         0    8184 10.1.1.145:80            220.160.210.6:52515          LAST_ACK     
tcp         1 14401 10.1.1.145:80            220.160.210.6:52514          LAST_ACK     
tcp         0    8184 10.1.1.145:80            220.160.210.6:52781          LAST_ACK     
是TCP80端口的应用,调节一下nginx的keepalive时间…
[root@ccsafe ~]# /usr/local/nginx/sbin/nginx -t -c /usr/local/nginx/conf/nginx.conf
2008/10/21 19:15:31 [info] 21352#0: the configuration file /usr/local/nginx/conf/nginx.conf syntax is ok
2008/10/21 19:15:31 [info] 21352#0: the configuration file /usr/local/nginx/conf/nginx.conf was tested successfully
[root@ccsafe ~]# ps aux|egrep ‘(PID|nginx)’
USER        PID %CPU %MEM     VSZ    RSS TTY       STAT START    TIME COMMAND
root       8290 0.0 0.0    7572 1124 ?         Ss    Oct04    0:00 nginx: master process /usr/local/nginx/sbin/nginx
nobody     8291 0.2 0.3 19704 13776 ?         S     Oct04 71:35 nginx: worker process      
nobody     8292 0.3 0.2 17604 11680 ?         S     Oct04 77:26 nginx: worker process      
nobody     8293 0.2 0.4 22528 16636 ?         S     Oct04 58:13 nginx: worker process      
nobody     8294 0.3 0.4 24944 19020 ?         S     Oct04 94:07 nginx: worker process      
nobody     8295 0.3 0.5 27496 21508 ?         S     Oct04 84:41 nginx: worker process      
nobody     8296 0.3 0.1 13388 7496 ?         S     Oct04 84:14 nginx: worker process      
nobody     8297 0.2 0.0    9196 3268 ?         S     Oct04 58:21 nginx: worker process      
nobody     8298 0.3 0.2 15392 9504 ?         S     Oct04 75:16 nginx: worker process      
root      21354 0.0 0.0    3896    720 pts/0     S+    19:15    0:00 egrep (PID|nginx)
(动态加载新配置)
[root@ccsafe ~]# kill -HUP 8290
[root@ccsafe ~]#
Every 10.0s: netstat -ant|fgrep :|cut -b 77-90 |sort |uniq -c                                                
       1 CLOSE_WAIT
    1138 CLOSING
    7161 ESTABLISHED
    1427 FIN_WAIT1
     396 FIN_WAIT2
    5740 LAST_ACK
       2 LISTEN
     350 SYN_RECV
     148 TIME_WAIT

[root@ccsafe ~]# netstat -ant|fgrep ":"|cut -b 77-90 |sort |uniq -c
    1151 CLOSING      
    8506 ESTABLISHED
    1452 FIN_WAIT1   
     666 FIN_WAIT2   
    6568 LAST_ACK     
       2 LISTEN      
     429 SYN_RECV     
      92 TIME_WAIT   

LAST_ACK不下,而且CLOSING 和FIN_WAIT突增
着重看看可影响主动断开TCP连接时几个参数
tcp_keepalive_intvl:探测消息发送的频率
tcp_keepalive_probes:TCP发送keepalive探测以确定该连接已经断开的次数
tcp_keepalive_time:当keepalive打开的情况下,TCP发送keepalive消息的频率
[root@ccsafe ~]# sysctl -a|grep tcp_keepalive
net.ipv4.tcp_keepalive_intvl = 30
net.ipv4.tcp_keepalive_probes = 2
net.ipv4.tcp_keepalive_time = 160
tcp_retries2:在丢弃激活(已建立通讯状况)的TCP连接之前?需要进行多少次重试
[root@ccsafe ~]# sysctl -a |grep tcp_retries
net.ipv4.tcp_retries2 = 15
net.ipv4.tcp_retries1 = 3
加速处理那些等待ACK的LAST_ACK,减少等待ACK的LAST_ACK的重试次数
[root@ccsafe ~]# sysctl -w net.ipv4.tcp_retries2=5
net.ipv4.tcp_retries2 = 5
减少keepalive发送的频率
[root@ccsafe ~]# sysctl -w net.ipv4.tcp_keepalive_intvl=15
net.ipv4.tcp_keepalive_intvl = 15
[root@ccsafe ~]# sysctl -p
排除syncookies的影响
[root@ccsafe ~]# !ec
echo "0" >/proc/sys/net/ipv4/tcp_syncookies
[root@ccsafe ~]# echo "1" >/proc/sys/net/ipv4/tcp_syncookies
[root@ccsafe ~]# sysctl -a|grep tcp_keepalive               
net.ipv4.tcp_keepalive_intvl = 30
net.ipv4.tcp_keepalive_probes = 2
net.ipv4.tcp_keepalive_time = 160
[root@ccsafe ~]# sysctl -a|grep syncookies
net.ipv4.tcp_syncookies = 1
延长keepalive检测周期,保留ESTABLISHED数量
[root@ccsafe ~]# echo "1800" >/proc/sys/net/ipv4/tcp_keepalive_time
[root@ccsafe ~]# echo "5" >/proc/sys/net/ipv4/tcp_keepalive_probes
[root@ccsafe ~]# echo "15" >/proc/sys/net/ipv4/tcp_keepalive_intvl
[root@ccsafe ~]# sysctl -a|grep tcp_keepalive                       
net.ipv4.tcp_keepalive_intvl = 15
net.ipv4.tcp_keepalive_probes = 5
net.ipv4.tcp_keepalive_time = 1800
[root@ccsafe ~]# !wat
watch -n 10 "netstat -ant|fgrep ":"|cut -b 77-90|sort |uniq -c"
Every 10.0s: netstat -ant|fgrep :|cut -b 77-90|sort |uniq -c                          
       1 CLOSE_WAIT
     363 CLOSING
    5145 ESTABLISHED
    1073 FIN_WAIT1
     174 FIN_WAIT2
    6042 LAST_ACK
       2 LISTEN
     301 SYN_RECV
      85 TIME_WAIT
LAST_ACK不下,但是CLOSING有所回落
tcp_orphan_retries:在近端丢弃TCP连接之前?要进行多少次重试。
[root@ccsafe ~]# sysctl -a|grep tcp_orphan
net.ipv4.tcp_orphan_retries = 0
关键,丢TCP太频繁了,以至于后勤都跟不上。设置丢弃之前的重试次数
[root@ccsafe ~]# echo "3" >/proc/sys/net/ipv4/tcp_orphan_retries
[root@ccsafe ~]# !wat
watch -n 10 "netstat -ant|fgrep ":"|cut -b 77-90|sort |uniq -c"
Every 10.0s: netstat -ant|fgrep :|cut -b 77-90|sort |uniq -c                          
       1 CLOSE_WAIT
      24 CLOSING
    5422 ESTABLISHED
     279 FIN_WAIT1
     214 FIN_WAIT2
    1966 LAST_ACK
       2 LISTEN
     269 SYN_RECV
      74 TIME_WAIT
上下调节该值,找个合适的临界点
[root@ccsafe ~]# echo "7" >/proc/sys/net/ipv4/tcp_orphan_retries                  
[root@ccsafe ~]# !wat
watch -n 10 "netstat -ant|fgrep ":"|cut -b 77-90|sort |uniq -c"
Every 10.0s: netstat -ant|fgrep :|cut -b 77-90|sort |uniq -c                          
       1 CLOSE_WAIT
     175 CLOSING
    5373 ESTABLISHED
     436 FIN_WAIT1
     209 FIN_WAIT2
    3184 LAST_ACK
       2 LISTEN
     283 SYN_RECV
     110 TIME_WAIT
恢复,同时FIN_WAIT1的值过高。考虑减少tcp_fin_timeout时间
[root@ccsafe ~]# echo "2" >/proc/sys/net/ipv4/tcp_orphan_retries                  
[root@ccsafe ~]# sysctl -a|grep tcp_fin
net.ipv4.tcp_fin_timeout = 10
[root@ccsafe ~]# echo "5" >/proc/sys/net/ipv4/tcp_fin_timeout
[root@ccsafe ~]# !wat
watch -n 10 "netstat -ant|fgrep ":"|cut -b 77-90|sort |uniq -c"
Every 10.0s: netstat -ant|fgrep :|cut -b 77-90|sort |uniq -c                          
       2 CLOSE_WAIT
      17 CLOSING
    5665 ESTABLISHED
     145 FIN_WAIT1
     141 FIN_WAIT2
    1068 LAST_ACK
       2 LISTEN
     287 SYN_RECV
      68 TIME_WAIT
相比FIN_WAIT,SYN_RECV的值偏高。加大发送synack的质量
[root@ccsafe ~]# sysctl -a|grep synack
net.ipv4.tcp_synack_retries = 1
[root@ccsafe ~]# echo "2" >/proc/sys/net/ipv4/tcp_synack_retries
[root@ccsafe ~]# !wat
watch -n 10 "netstat -ant|fgrep ":"|cut -b 77-90|sort |uniq -c"
Every 10.0s: netstat -ant|fgrep :|cut -b 77-90|sort |uniq -c                          
       3 CLOSE_WAIT
      16 CLOSING
    5317 ESTABLISHED
     200 FIN_WAIT1
     158 FIN_WAIT2
    1001 LAST_ACK
       2 LISTEN
     303 SYN_RECV
      78 TIME_WAIT
[root@ccsafe ~]# sysctl -a|grep keepalive
net.ipv4.tcp_keepalive_intvl = 15
net.ipv4.tcp_keepalive_probes = 5
net.ipv4.tcp_keepalive_time = 1800
[root@ccsafe ~]# watch -n 10 "netstat -ant|fgrep ":"|cut -b 77-90|sort |uniq -c"
Every 10.0s: netstat -ant|fgrep :|cut -b 77-90|sort |uniq -c                          
       1 CLOSE_WAIT
       7 CLOSING
    5356 ESTABLISHED
     175 FIN_WAIT1
     136 FIN_WAIT2
    1045 LAST_ACK
       2 LISTEN
     345 SYN_RECV
      64 TIME_WAIT
减少keepalive的检测周期,LAST_ACK上升
[root@ccsafe ~]# echo "10" >/proc/sys/net/ipv4/tcp_keepalive_intvl
[root@ccsafe ~]# echo "1" >/proc/sys/net/ipv4/tcp_synack_retries                  
[root@ccsafe ~]# !wat
watch -n 10 "netstat -ant|fgrep ":"|cut -b 77-90|sort |uniq -c"
Every 10.0s: netstat -ant|fgrep :|cut -b 77-90|sort |uniq -c                                                
       1 CLOSE_WAIT
      13 CLOSING
    5605 ESTABLISHED
     212 FIN_WAIT1
     131 FIN_WAIT2
    1143 LAST_ACK
       2 LISTEN
     252 SYN_RECV
      79 TIME_WAIT
恢复
[root@ccsafe ~]# echo "15" >/proc/sys/net/ipv4/tcp_keepalive_intvl               
[root@ccsafe ~]# watch -n 10 "netstat -ant|fgrep ":"|cut -b 77-90|sort |uniq -c"
Every 10.0s: netstat -ant|fgrep :|cut -b 77-90|sort |uniq -c                        
       3 CLOSE_WAIT
      14 CLOSING
    5862 ESTABLISHED
     230 FIN_WAIT1
     205 FIN_WAIT2
    1064 LAST_ACK
       2 LISTEN
     244 SYN_RECV
      59 TIME_WAIT
[root@ccsafe ~]# watch -n 10 "netstat -ant|fgrep ":"|cut -b 77-90|sort |uniq -c"
Every 10.0s: netstat -ant|fgrep :|cut -b 77-90|sort |uniq -c                        
       3 CLOSE_WAIT
      26 CLOSING
    6712 ESTABLISHED
     270 FIN_WAIT1
     230 FIN_WAIT2
     994 LAST_ACK
       2 LISTEN
     254 SYN_RECV
      73 TIME_WAIT
[root@ccsafe ~]#
目前LAST_ACK占ESTABLISHED的量在15%左右
文章来源: http://hi.baidu.com/_wang8/blog/ … 6a8b370b55a971.html

文章出自:CCIE那点事 http://www.jdccie.com/ 版权所有。本站文章除注明出处外,皆为作者原创文章,可自由引用,但请注明来源。 禁止全文转载。
本文链接:http://www.jdccie.com/?p=1510转载请注明转自CCIE那点事
如果喜欢:点此订阅本站
  • 相关文章
  • 为您推荐
  • 各种观点

暂时还木有人评论,坐等沙发!
发表评论

您必须 [ 登录 ] 才能发表留言!