CPU utilization is high on AIX node1 in the RAC
CPU utilization is high on AIX node1 in the RAC.
Alert : EM Event: Warning:ORAX01 – CPU Utilization is 82.927%, crossed warning (80) or critical (95) threshold.
Issue:
CPU utilization is high on AIX node1 in the RAC due to few sessions.
Root cause:
The load on the server causing by , Java, Asm process and DBSNMP processes.
Action Taken:
Restarted the Agent and cleaned dbsnmp sessions in DB level.
The load on the server has come down from 33 to 10%
Load Average Before :
oraax01:/u00/app/oracle:SID=OPRD1 > uptime
04:38PM up 431 days, 3:52, 2 users, load average: 38.04, 31.13, 29.85
Load Average After:
oraax01:/u00/app/oracle:SID=OPRD1 > uptime
05:07PM up 431 days, 4:20, 3 users, load average: 10.78, 12.00, 18.28
PID Details
PID that causing high load on the server
One is Java(OEM agent), one is ASM process and the remaining process are DBSNMP processes of each database running on this cluster.
topas command on AIX Node1
Topas Monitor for host:oraax01
CPU User% Kern% Wait% Idle% Physc Entc%
Total 62.4 12.1 0.2 25.3 0.99 99.13
Name PID CPU% PgSp Owner PageIn 0 PAGING SPACE
java 11862076 24.4 485M oracle PageOut 446 Size,MB 40192
lsvpd 4063650 14.8 1.16M root Sios 446 % Used 0
oracle 15073418 13.1 1.34G oracle % Free 100
oracle 24510540 12.0 307M oracle NFS (calls/sec)
oracle 42729488 11.3 307M oracle
sample alert message from OEM server looks like below
EM Event: Warning:ORAX01 – CPU Utilization is 82.927%, crossed warning (80) or critical (95) threshold.
Host=ORAX01
Target type=Host
Target name=ORAX01
Categories=Capacity
Message=CPU Utilization is 82.927%, crossed warning (80) or critical (95) threshold.
Severity=Warning
Operating System=AIX
Platform=powerpc
Event Type=Metric Alert
Event name=LPAR Performance on AIX:pcpuUtil
Metric Group=LPAR Performance on AIX
Metric=CPU Utilization (%)
Metric value=82.927
Key Value=
Rule Name=Incident management rule set for all targets,Clear metric alert events older than 7 days
Rule Owner=System Generated
See Also: