Forum Discussion
cathy_123
Cirrus
CirrusHigh CPU Utilization CPU1
Hello F5!
Would like to ask for possible cause and fix. We have high CPU utlization on CPU1 . It seems it is the systems causing our CPU1 high utilization. Our F5 is in Virtual and in version 12.1.2. Thank you!
CPU: 0 (clock ticks) Last 5 sec Last 1 min Last 5 min Total
- (avg/sec) (avg/sec) (avg/sec) -
User 8 3 2 26.7M
Niced 0 0 0 3.9M
System 58 13 3 17.3M
Idle 33 83 94 2.7G
Irq 0 0 0 8.5K
Softirq 0 0 0 2.0M
Iowait 0 0 0 2.0M
Stolen 0 0 0 0
Util% (last 10 sec) - - - 0
CPU: 1 (clock ticks) Last 5 sec Last 1 min Last 5 min Total
- (avg/sec) (avg/sec) (avg/sec) -
User 5 2 1 28.3M
Niced 0 0 0 6.5M
System 30 4 1 45.6M
Idle 63 94 97 2.7G
Irq 0 0 0 582
Softirq 0 0 0 501.0K
Iowait 0 0 0 2.5M
Stolen 0 0 0 0
Util% (last 10 sec) - - - 100
HamishCirrocumulus
Have a look and see which process is using all that CPU...
'top' may help there...
H
Hi Cathy
There is no high CPU utilisation from the output you posted.
The Idle output is just how much of its time the CPU was just sitting there doing nothing.
The System output average for the past 5 minutes for CPU 1 was just 1%.
The rest is all zeroes.
cathy_123Thanks Everyone!
Hi Hamish I did try top command see below
Swap: 1023996k total, 782240k used, 241756k free, 262580k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
1 root 20 0 598m 188m 1140 R 96.6 4.8 8122:08 init . -> init has 96% utilization
10580 root RT 0 2790m 71m 52m S 3.9 1.8 6555:32 tmm.0
9 root 20 0 0 0 0 S 2.0 0.0 74:45.46 ksoftirqd/1
5796 root 20 0 44756 17m 15m S 2.0 0.5 1427:46 bigd
Hi Rodrigo,
Would like to ask isnt the CPU0 has 0% Util and CPU1 has 100% Util?
Thank you!
ah, yeah. I missed that.. sorry.
Are you able to take a snapshot of your system and upload to ihealth? https://ihealth.f5.com
You can log in to BIG-IP and click on System --> Support.
You can then upload your snapshot (called qkview) to ihealth and lastly copy the ihealth link and paste it to us.
I can have a quick look at your CPU usage.
- cathy_123
Thanks Rodrigo I did upload it. will you be able to see this https://ihealth.f5.com/qkview-analyzer/qv/11773442/status/overview?offset=0 thanks
- Hamish
96% utilisation for init isn't normal.
It's a centos (i.e. RedHat) base. (Not sure which one for 12.1.2 - But my v24 uses Centos 6.8).
However it shouldn't normally use a lot of CPU. Anything logged? (/var/adm/messages) or in the 'dmesg' output?
Have you edited inittab at all? Normally there's just a single entry in there, but some people (Myself also) like to put important daemons in there with restart enabled. Breaking those can cause init to spin.
I take it from the discrepenecy between your tmmsh cpu output and top that it's bursty?
H
