Shell script to handle Nginx site logs

In the Nginx site, access and error are using the default log format, with the log file named as follows:

Www.xxxx.com.access.log

Www.xxxx.com.error.log

The script processes the previous day's log records, re-places the logs in compressed files, and stores them by month, as follows:

201611/www.xxxx.com.access.20161102.log.zip

201611/www.xxxx.com.error.20161102.log.zip

The script is as follows:

#!/bin/bash# description: deal nginx access and error log dailyyes_day _a= ' date +%d/%b/%y -d -1day ' yes_day_a1= ' date +%d '/'%b ' \ '%y -d -1day ' yes_day_E= ' Date +%y/%m/%d -d -1day ' yes_day_e1= ' date +%y '/'%m ' \ '%d -d -1day ' yes_day= ' date  +%y%m%d -d -1day ' mon= ' date +%y%m ' file_list= ' find /etc/nginx/ -type f  -name  "*.conf"  -print0 | xargs -0 egrep  ' ^ (\s|\t) *access_log|error_log ' |awk   ' {print $3} ' |tr  '; "   " " |tr  ' \ n '   '   ' for file in  $file _listdo         site= ' echo  $file  |awk -f "log$"   ' {print $1} '          count=0        if [ -f   $file  ];THEN            &Nbsp;   echo  "Log file found: $file"                  path= ' dirname  $file '                  cd  $path  && find  ./ -name  "*[0-9].log.zip"  -mtime +30 -delete                 test ! -d  $mon  & & mkdir  $mon                                  if  [[  $file  =~ .*acc.* ]];then                         date_fmt= "\[$yes _day _a: "                         date_str= "\[$yes _day_a1:"                          echo -e  "\taccess_file: $file"                           #continue                  elif [[  $file  =~ .*err.* ]];then                          date_fmt= "^ $yes _day_e"                          date_str= "^ $yes _day_e1"                          echo -e  "\terror_file: $file"                           #continue                  fi                                  dst_file= "${site}${yes_day}.log"                  grep  "$date _fmt"   $file  >$ dst_file                [  $? -ne 0 ] && let count=count+1                 sleep 5                                  dfile_size= ' ls -l  $dst _file |awk  ' {print $5} '                  if [  $dfile _size  -gt 1024 ];then                         zip -m -r ${dst_file}.zip   $DST _file                         mv ${dst_file}.zip  $mon/                 elif [  $dfile _size  -le 10 ];then                         rm -f  $DST _file                 else                         mv  $DST _file $ mon/                fi                                [ $? -ne 0 ] & & let count=count+1                 sleep 5                                 if [  $count  -eq 0 ];then                         sed -i  "/$date _str/d"   $file                          sed -i  "/^$/d"   $file                  fi         Fidonenginx -t[ $? -eq 0 ] && nginx -s reload

Where a pit is encountered: the SED command deletes the file, the content that is about to be deleted is temporarily placed in the current directory in the file that begins with SED, the contents of the source file are deleted, and the temporary file is finally deleted. If an exception occurs in the middle (the temporary file causes insufficient disk space), the delete operation is terminated, the source file is not changed, and the temporary file is not processed automatically.

This article from the "11414123" blog, reproduced please contact the author!

Shell script to handle Nginx site logs