1.定义nginx日志格式
log_format main '$remote_addr - [$time_local] "$request" '
' - $status "User_Cookie:$http_cookie" ';
2.定义access日志的输出位置和调用哪个日志模板
access_log logs/access.log main;
日志内容是这样的，最后会显示用户的cookie
192.168.40.2 - [02/Nov/2016:15:44:35 +0800] "GET /wcm/app/main/refresh.jsp?r=1478072325778 HTTP/1.1" - 200 "User_Cookie:7F00000122A5597C46607B1C0A7EC016" 192.168.40.2 - [02/Nov/2016:15:44:35 +0800] "GET /webpic/W0201611/W020161102/W020161102566715167404.jpg HTTP/1.1" - 200 "User_Cookie:7F00000122A5597C46607B1C0A7EC016" 119.255.31.109 - [02/Nov/2016:15:44:36 +0800] "GET /wcm/app/main/refresh.jsp?r=1478072510132 HTTP/1.1" - 200 "User_Cookie:7F000001237921BE9237838AEC65704D" 119.255.31.109 - [02/Nov/2016:15:44:36 +0800] "GET /wcm/app/message/message_query_service.jsp?READFLAG=0&MSGTYPES=1%2C2%2C3 HTTP/1.1" - 200 "User_Cookie:7F000001237921BE9237838AEC65704D" 192.168.40.2 - [02/Nov/2016:15:44:37 +0800] "GET /wcm/app/message/message_query_service.jsp?READFLAG=0&MSGTYPES=1%2C2%2C3 HTTP/1.1" - 200 "User_Cookie:7F00000123D3BF2345115EAAC21F71E0" 192.168.40.2 - [02/Nov/2016:15:44:37 +0800] "GET /wcm/app/message/message_query_service.jsp?READFLAG=0&MSGTYPES=1%2C2%2C3 HTTP/1.1" - 200 "User_Cookie:7F00000123EF73896DF98EDA9950944E" 192.168.40.2 - [02/Nov/2016:15:44:37 +0800] "GET /wcm/app/message/message_query_service.jsp?READFLAG=0&MSGTYPES=1%2C2%2C3 HTTP/1.1" - 200 "User_Cookie:7F00000123FE0F9C397E1A8F0C4F044B" 192.168.40.2 - [02/Nov/2016:15:44:37 +0800] "GET /wcm/app/main/refresh.jsp?r=1478072511427 HTTP/1.1" - 200 "User_Cookie:7F00000123A465B7EA1DE0AF0AE671B7" 119.255.31.109 - [02/Nov/2016:15:44:38 +0800] "GET /wcm/app/message/message_query_service.jsp?READFLAG=0&MSGTYPES=1%2C2%2C3 HTTP/1.1" - 200 "User_Cookie:7F00000123D89B11302DF80AE773C900"
接下来的思路就很简单，如果统计pv，就直接统计日志的行数，如果统计UV，就根据cookie去重。
PV统计
可统计单个链接地址访问量：
grep index.shtml host.access.log | wc -l
总PV量：
awk '{print $6}' host.access.log | wc -l
独立IP：
awk '{print $1}' host.access.log | sort -r |uniq -c | wc -l
UV统计：
awk '{print $10}' host.access.log | sort -r |uniq -c |wc -l