Shell - sar 기반 시간대별 load average 간단 추출 및 확인

Linux OS는 기본적으로 /var/log/sa 이하에 sar 성능 raw 데이터를 적재하고 있다.

 

시스템의 장애시나 성능지연등의 이상징후시 해당 RAW 데이터를 활용하여 분석에 활용 가능하지만,

 

sar 커맨드나 ksar등의 특성상 전체 일자 기준이 아닌 각각의 일자 데이터를 따로 개별 확인을 해야한다.

 

아래 Script를 활용하면 /var/log/sa 이하에 적재된 sar 데이터를 기준으로 각 일자별, 시간대별 Load Average를 손쉽게 확인이 가능하다.

 

[ 기준 데이터 ]   [root@TestVM01 ~]# [root@TestVM01 ~]# ls -l /var/log/sa/sa[0-9]* -rw-r--r-- 1 root root  352752 Dec  1 23:50 /var/log/sa/sa01 -rw-r--r-- 1 root root  352752 Dec  2 23:50 /var/log/sa/sa02 -rw-r--r-- 1 root root  352752 Dec  3 23:50 /var/log/sa/sa03 -rw-r--r-- 1 root root  352752 Dec  4 23:50 /var/log/sa/sa04 -rw-r--r-- 1 root root 2044784 Dec  5 23:59 /var/log/sa/sa05 -rw-r--r-- 1 root root 3525360 Dec  6 23:59 /var/log/sa/sa06 -rw-r--r-- 1 root root 3525360 Dec  7 23:59 /var/log/sa/sa07 -rw-r--r-- 1 root root 2962320 Dec  8 20:09 /var/log/sa/sa08 [root@TestVM01 ~]#

 

물론 아래 Script 이하 sar command 부분을 개작하면 Load Average 뿐만이 아닌 CPU, I/O 수치등도 원하는데로 추출이 가능하다.

 

 

1. 사용법

 

[root@TestVM01 ~]# [root@TestVM01 ~]# ./load_max.sh Usage1 - ./load_max.sh[Start Time] [End Time] ex1) - ./load_max.sh07:30:00 10:00:00 [root@TestVM01 ~]# [root@TestVM01 ~]# [root@TestVM01 ~]# ./load_max.sh 07:30:00 10:00:00 [ Load Average - Dec 1 - /var/log/sa/sa01 ] TIME=07   Average=0.95   Min=0.73   Max=1.18 TIME=08   Average=5.16   Min=0.77   Max=23.78 TIME=09   Average=3.36   Min=0.77   Max=8.55 #################################################################### [ Load Average - Dec 2 - /var/log/sa/sa02 ] TIME=07   Average=1.68   Min=0.77   Max=2.24 TIME=08   Average=5.13   Min=0.77   Max=9.62 TIME=09   Average=11.45   Min=0.77   Max=23.19 #################################################################### [ Load Average - Dec 3 - /var/log/sa/sa03 ] TIME=07   Average=0.99   Min=0.49   Max=1.50 TIME=08   Average=1.62   Min=0.45   Max=3.67 TIME=09   Average=10.99   Min=0.45   Max=15.98 #################################################################### [ Load Average - Dec 4 - /var/log/sa/sa04 ] TIME=07   Average=1.45   Min=0.76   Max=2.14 TIME=08   Average=5.38   Min=0.86   Max=18.08 TIME=09   Average=6.87   Min=0.86   Max=12.49 #################################################################### [ Load Average - Dec 5 - /var/log/sa/sa05 ] TIME=07   Average=1.19   Min=0.33   Max=2.04 TIME=08   Average=1.67   Min=0.85   Max=3.43 TIME=09   Average=335.95   Min=0.85   Max=1317.05 #################################################################### [ Load Average - Dec 6 - /var/log/sa/sa06 ] TIME=07   Average=0.71   Min=0.53   Max=2.00 TIME=08   Average=0.73   Min=0.60   Max=5.87 TIME=09   Average=1.74   Min=0.76   Max=21.01 #################################################################### [ Load Average - Dec 7 - /var/log/sa/sa07 ] TIME=07   Average=0.59   Min=0.19   Max=1.59 TIME=08   Average=0.73   Min=0.79   Max=2.18 TIME=09   Average=0.98   Min=0.72   Max=3.52 #################################################################### [ Load Average - Dec 8 - /var/log/sa/sa08 ] TIME=07   Average=1.07   Min=0.37   Max=3.15 TIME=08   Average=4.74   Min=0.93   Max=34.69 TIME=09   Average=11.64   Min=0.93   Max=40.03 #################################################################### [root@TestVM01 ~]#

 

 

2. Script 내용

 

[root@TestVM01 ~]# [root@TestVM01 ~]# vi ./load_max.sh #!/bin/bash ##By Kwang-min.choi@kt.com   export LANG=C export LC_ALL=C   if [ $# -ne 2 ] then  echo  echo "Usage1 - $0 [Start Time] [End Time]"  echo "ex1) - $0 07:30:00 10:00:00"  echo  exit 0 fi   for SAR_LIST in `ls -1rt /var/log/sa/sa[0-9]*` do  SAR_NAME=`echo ${SAR_LIST} | cut -d "/" -f5`  sar -q -f ${SAR_LIST} -s $1 -e $2 | awk '$1 ~ /^[0-9]/ && $2 ~ /^[0-9]/ {print $1, $4}' > ${SAR_NAME}.raw  echo  echo "[ Load Average - `ls -l ${SAR_LIST} | awk '{print $6, $7}'` - ${SAR_LIST} ]"  echo  RAWDATA=${SAR_NAME}.raw  cat ./${RAWDATA} | cut -d ":" -f1 | uniq > uniq.list    while read UNIQ_LIST;  do   TOTAL_SUM=0   TOTAL_LIST=`cat ${RAWDATA} | sed 's#%##g' | grep "^${UNIQ_LIST}"`   TOTAL_RAW=`cat ${RAWDATA} | sed 's#%##g' | grep "^${UNIQ_LIST}" | awk '{print $2}'`   TOTAL_COUNT=`cat ${RAWDATA} | sed 's#%##g' | grep "^${UNIQ_LIST}" | wc -l`   MAX=0     for LIST in ${TOTAL_RAW}   do    export TOTAL_SUM=`echo "${TOTAL_SUM} + ${LIST}" | bc`    export DIFF=`echo "${LIST} > ${MAX}" | bc`    if [ "$DIFF" -eq 1 ];    then     MAX=${LIST}    fi      if [ -z "$MIN" -a -z "$BEFOR_LIST" ];    then     MIN=${LIST}     BEFOR_LIST=${LIST}    else     export MIN_DIFF=`echo "${LIST} <= ${BEFOR_LIST}" | bc`       if [ "$MIN_DIFF" -eq 1 ];     then      MIN=${LIST}     fi    fi    done   export AVERAGE=`echo "${TOTAL_SUM} ${TOTAL_COUNT}" | awk '{printf "%.2f", $1 / $2}'`     echo -e "TIME=$UNIQ_LIST \t Average=${AVERAGE} \t Min=${MIN} \t Max=${MAX}"  done < uniq.list echo echo "####################################################################" done