서버의 load average가 1.02 이상 올라가는 상황 발생
조치 후에 load average 0.43 이하로 감소된 것 확인
문제가 되었던 것은 kipmi0 라는 데몬이 CPU 증가를 야기하였음.
OS 환경 = CentOS release 6.5 (Final) x86_64
증상
PID USER PR NI VIRT RES SHR S %CPU TIME %MEM TIME+ COMMAND
130 root 39 19 0 0 0 R 99.5 1074h 0.0 64442:54 kipmi0
해결 방법
참고 URL : http://portal.solutionengineers.net/knowledgebase.php?action=displayarticle&id=6
Seeing 1.00 average server load – 100% of a single core used by kipmi0.
Resolved by executing from shell, as root –
echo 100 > /sys/module/ipmi_si/parameters/kipmid_max_busy_us
Another note:
On CentOS/RHEL, that command can be made permanent by adding it to /etc/rc.d/rc.local. The rc.local runs after all of the other init scripts
***************************************************************************
적용
# echo 100 > /sys/module/ipmi_si/parameters/kipmid_max_busy_us
리부팅 후에도 적용되길 원한다면
/etc/rc.d/rc.local 에 해당 명령어 기록
적용 확인
PID USER PR NI VIRT RES SHR S %CPU TIME %MEM TIME+ COMMAND
130 root 39 19 0 0 0 S 10.9 1074h 0.0 64443:07 kipmi0
[polldaddy rating=”7739789″]