For more information regarding the security incident at F5, the actions we are taking to address it, and our ongoing efforts to protect our customers, click here.

Forum Discussion

SanYang's avatar
SanYang
Icon for Cirrostratus rankCirrostratus
Sep 14, 2023

High CPU utilization for only cpu 0

Hello, I have 26 CPUs in multi thread format but only CPU 0 utilization goes 100 and other CPUs utilization remains under 30%. # tmsh show /sys tmm-info -------------------------- Sys::TMM: 0.0 ...
application delivery
Mohamed_Ahmed_Kansoh's avatar
Mohamed_Ahmed_Kansoh
Icon for MVP rankMVP
Sep 17, 2023

Hi SanYang, 

As I don't see errors in passing traffic and somehow balanced traffic for all TMM instances ( maybe only TMM0.10 handles high connections so it's CPU is +60% ) but over all TMMs CPU is within 20%. 

So this is an indication of that TMM0 represents the Total CPU & memory usage in your blade and because you have till 24 tmm instances it gave you 100 % as this is a high number of TMMs. 
This is Clarified in this Article : https://my.f5.com/manage/s/article/K02895424 

Now , you should keen on in the overall CPU & memory usage as average between all TMMs cores. 

If you go to ( Statistics >>> Performance reports >>> Chech the CPU & memory graphs ) 
Also adjust time range ( Last 3 hours then Last week , then last month ) 
I'm sure you will find the overall Cpu is consistent following the same trend for all other graphs , I mean you will find the overall/average CPU for the ( last month/week/day/hours ) not high or reached to 100% 

Do u know if you found this overall CPU reached to 100 % your box will failover immediately. 

So I need you to check that performance report and let me know if my analysis is correct or not. 

Thanks

SanYang's avatar
SanYang
Icon for Cirrostratus rankCirrostratus
Sep 17, 2023

Hi Mohamed_Ahmed_Kansoh ,

Thanks for your reply.

I checked the charts.
CPU0 has been running at 100% for a while now...

 

 

  • Mohamed_Ahmed_Kansoh's avatar
    Mohamed_Ahmed_Kansoh
    Icon for MVP rankMVP
    Sep 18, 2023

    Hello SanYang, 

    yes this does make sense with the TMM0 CPU value in CLI. 

    But as I see in the historical graphs last monday and tuesday it was normal. 

    I see now you need to open a case with F5 support , to validate the health of TMM0. 
    maybe this is the reason : https://my.f5.com/manage/s/article/K02895424

    but you need to verify this , because if one of your TMMs crashes ( data plane ones I mean ) this will effect badly in your overall performance. 

    So I recommend opening a case. 

    then when you get the RCA for this high utilization you can share it with us.