shell脚本：自动内存监控及日志备份

短命女 2022-08-24 01:58 235阅读 0赞

由于系统经常出现内存溢出的情况，需要对系统内存的实时情况进行监控，一旦发现大于指定的临界值时，根据指定邮件列表发送邮件，自动**备份**并压缩日志，上传到指定的ftp空间上。  
并且实现了对于单个进程只做单次处理，只有在重启之后，才会重新的新进程进行响应（内存超标的时候）。

以下的脚本实现了这个功能，自动ftp上传的功能在另外一个脚本中实现(20080620更新新版本)：  
新版本：  
说明：由于发现执行kill -3时可能对tomcat本身的**java**进程产生影响，使内存升高加剧，所以增加了2此检查功能，在达到1500M时，发送邮件提醒，但不做kill -3。因为此时可能内存还有降低的情况存在。一直到2000M时，才执行kill-3。  
\--1.增加了两次检查功能  
\#!/bin/sh

USER="regonly1";  
CHKFLAG="java";  
curdir=\`pwd\`;  
runlogdir="$HOME/runlog";  
msglogdir="$HOME/log";  
\#runlogdir="$HOME/local"; \# Only for test.  
maxchksize=1500; \# If memory size great than maxchksize, mail to the users according to the mail list.  
maxkill3sz=2000; \# When memory size great than maxkill3sz, backup files and execute kill -3;  
checkiter=1;  
oldprocid=0;  
interval=8;  
\#iter=0;

\#Mail list. Use "," to delimit every mail address.  
mailto="[regonly1@abc.com][regonly1_abc.com]";

filetrans()\{  
 backupdate=\`date +%Y%m%d\`;  
 if test -f $\{curdir\}/autoftptrans.epc  
 then  
  $\{curdir\}/autoftptrans.epc "$1" "$\{backupdate\}";  
 else  
  echo "Request file dose not exist: $\{curdir\}/autoftptrans.epc.";  
 fi;  
\}  
packmsglogfile()\{  
 echo "Begin backup message file logs....";  
 cd $\{msglogdir\};  
 mkdir temp;  
 targetfile=MsgFile.$\{chkdate\}.tar.gz;  
 for f in \`ls Message\*\`; do  
  cp $f temp/ ;  
 done  
 tar zcvf $\{targetfile\} temp;  
 filetrans "$\{targetfile\}";  
 rm -fr temp "$\{targetfile\}";  
 echo "MessageFileLogs backup OK.";  
\}  
packrunlogfile()\{  
 echo "Begin backup runlogs....";  
 cd $runlogdir;  
 cd \`find ./ -type d | ls -lrt | tail -1 | awk '\{print $9\}'\`;  
 curfile=\`find ./ -type f | ls -lrt | tail -1 | awk '\{print $9\}'\`;  
 targetfile="catalina.$\{chkdate\}.tar.gz";  
 cp $\{curfile\} $\{curfile\}.kk;  
 tar zcvf $\{targetfile\} $\{curfile\}.kk;  
 filetrans "$\{targetfile\}";  
 rm -fr $\{curfile\}.kk $\{targetfile\};  
 echo "Runlog backup OK!";  
\}

while \[ 1 \]; do  
 getproc=\`top -b -U $\{USER\} | grep $\{USER\} | grep $\{CHKFLAG\}\`;  
 memsize=\`echo $getproc | awk '\{print substr($7,1,length($7)-1);\}'\`;  
 cpupcnt=\`echo $getproc | awk '\{print $10;\}'\`;  
 procsid=\`echo $getproc | awk '\{print $1;\}'\`;  
 chktime=\`date +"%Y-%m-%d %H:%M:%S"\`;  
 chkdate=\`date +"%Y%m%d"\`;  
 if \[ "$\{memsize\}" -gt "$\{maxchksize\}" -a "$\{oldprocid\}" -ne "$\{procsid\}" \];then  
  oldprocid=$\{procsid\};  
      echo "Memory overflow: $\{memsize\}M";  
  echo "Send mail to " $\{mailto\};  
  echo "Check time:\[$\{chktime\}\]  Memory used:$\{memsize\}M. Host:\`hostname\`." | mail $\{mailto\};  
                checkiter=1;  
 elif \[ "$\{memsize\}" -gt "$\{maxchksize\}" -a "$\{oldprocid\}" -eq "$\{procsid\}" \]; then  
  if \[ "$\{memsize\}" -gt "$\{maxkill3sz\}" -a "$\{checkiter\}" -eq 1 \]; then  
   echo "Memory size large than $\{memsize\}.";  
   echo "Begin to backup logfiles...";  
   packmsglogfile;  
   packrunlogfile;  
   kill -3 $\{procsid\};  
   echo "Logfile backup OK!";  
   echo "Logfile backup OK." | mail $\{mailto\};  
                        checkiter=0;  
  fi;  
  echo "Memory overflow:$\{memsize\}M. CPU: $\{cpupcnt\}. Please check memory in time!";  
  sleep $\{interval\};  
 else  
  echo "Memory status OK. Memory: $\{memsize\}M. CPU: $\{cpupcnt\}. at \`hostname\` \[$\{chktime\}\].";  
  sleep $\{interval\};  
 fi;  
done;  
  
  
  
\--原来版本  
\#!/bin/sh

USER="regonly1";\#指定top的用户  
CHKFLAG="java";\#监测的进程标志  
curdir=\`pwd\`;  
runlogdir="$HOME/runlog";  
msglogdir="$HOME/log";  
\#runlogdir="$HOME/local"; \# 仅用于测试  
maxsize=1500; \#监控临界点，超过这个值则发送邮件给指定的邮件用户  
oldprocid=0;\#记录老的进程id  
interval=8;\#记录监测的间隔时间  
\#iter=0;\#迭代器，取消

\#邮件列表，以","分隔  
mailto="mymail@abc.com";

\#自动传输函数，调用expect脚本来实现自动ftp上传的功能：  
filetrans()\{  
 backupdate=\`date +%Y%m%d\`;  
 $\{curdir\}/autoftptrans.epc "$1" "$\{backupdate\}";  
\}  
packmsgfile()\{  
 echo "Begin backup message file logs....";  
 cd $\{msglogdir\};  
 mkdir temp;  
 targetfile=MsgFile.$\{chkdate\}.tar.gz;  
 for f in \`ls Message\*\`; do  
  cp $f temp/ ;  
 done  
 tar zcvf $\{targetfile\} temp;  
 filetrans "$\{targetfile\}";  
 rm -fr temp "$\{targetfile\}";  
 echo "MessageFileLogs backup OK.";  
\}  
packrunlogfile()\{  
 echo "Begin backup runlogs....";  
 cd $runlogdir;  
 cd \`find ./ -type d | ls -lrt | tail -1 | awk '\{print $9\}'\`;  
 curfile=\`find ./ -type f | ls -lrt | tail -1 | awk '\{print $9\}'\`;  
 targetfile="catalina.$\{chkdate\}.tar.gz";  
 cp $\{curfile\} $\{curfile\}.kk;  
 tar zcvf $\{targetfile\} $\{curfile\}.kk;  
 filetrans "$\{targetfile\}";  
 rm -fr $\{curfile\}.kk $\{targetfile\};  
 echo "Runlog backup OK!";  
\}

while \[ 1 \]; do  
 getproc=\`top -b -U $\{USER\} | grep $\{USER\} | grep $\{CHKFLAG\}\`;  
 memsize=\`echo $getproc | awk '\{print substr($7,1,length($7)-1);\}'\`;  
 cpupcnt=\`echo $getproc | awk '\{print $10;\}'\`;  
 procsid=\`echo $getproc | awk '\{print $1;\}'\`;  
 chktime=\`date +"%Y-%m-%d %H:%M:%S"\`;  
 chkdate=\`date +"%Y%m%d"\`;  
\#对于同一个进程id，只做一次内存超出邮件提醒，后面再进程重启之前都不会再对该进程监测  
 if \[ "$\{memsize\}" -gt "$\{maxsize\}" -a "$\{oldprocid\}" -ne "$\{procsid\}" \];then  
  oldprocid=$\{procsid\};  
      echo "Memory overflow: $\{memsize\}M";  
  echo "Send mail to " $\{mailto\};  
  echo "Check time:\[$\{chktime\}\]  Memory used:$\{memsize\}M. Host:\`hostname\`." | mail $\{mailto\};  
  kill -3 $\{procsid\};  
  packmsgfile;  
  packrunlogfile;  
 elsif \[ "$\{memsize\}" -gt "$\{maxsize\}" -a "$\{oldprocid\}" -eq "$\{procsid\}" \]  
  echo -n "Memory overflow:$\{memsize\}M. CPU: $\{cpupcnt\}.";  
  echo "Logs already been backuped. Please check memory in time!";  
  sleep $\{interval\};  
 else  
  echo "Memory status OK. Memory: $\{memsize\}M. CPU: $\{cpupcnt\}. at \`hostname\` \[$\{chktime\}\].";  
  sleep $\{interval\};  
 fi;  
done;

自动ftp上传脚本，其中文件参数和日期参数都是从上面的脚本中传过来的：  
\#!/usr/local/bin/expect

set file \[lindex $argv 0\];  
set date \[lindex $argv 1\];  
set timeout -1;  
while \{1\} \{  
 spawn ftp 192.168.11.24;  
 expect "Name";  
 send "ftpuser/r";  
 expect "Password";  
 send "ftppasswd/r";  
 expect "Login failed." \{send "quit/r";close;wait\} "ftp>" break;  
\}  
send "cd hzdtvoss/daycheck//r";  
expect "ftp>";  
send "mkdir $date/r";  
expect "ftp>";  
send "cd $date/r";  
expect "ftp>";  
send "bi/r";  
expect "ftp>";  
send "put $file/r"  
expect "ftp>";  
send "quit/r";

[regonly1_abc.com]: mailto:regonly1@abc.com