Forum Discussion

unavailable's avatar
unavailable
Icon for Cirrus rankCirrus
Jan 11, 2024

Overload Cpu

Hello everyone,

i have been migrating the configuration beetwen F5 device from old device to another one i2600 ver 16.1.4.

I have verified that the device does not responding as expected and i have verified by using graphic the use of it, in a short in some hours of the day it use  a lot of cpu and, after 06:00 p.m. the situation is normal.

You need to know thatm at the moment, the device does not processing traffic, i have configured only the high availability and management ip.

I have attached the creen about the use of the cpu and the script that are inclued in the crontab.

Someone could help me to fix the problem ?

Many thanks in advanced for your attention.

 

Rgds,

 

application delivery
overload cpu
  • PhatANhappy's avatar
    PhatANhappy
    Jan 11, 2024

    that could be the issue.   definatly something to look at.
    i would lean more towards using the top command to find out what is causing the issue.
    With ASM -  i would further be concerned about spinning disks vs. SSD which will show up as a  'wa' in the top stastics.

  • Nikoolayy1's avatar
    Nikoolayy1
    Jan 11, 2024

    PhatANhappy  mentioned about too small resources and that this is control plane utlization not data plane and the ''top'' will review if it is the bug or another process. About the ''wa'' that is something I would also renember from this discussion.

     

    About the Analysis plane this in most cases is the ASM/AWAF policy builder learning, so also better also switch it off The BIG-IP ASM device shows high CPU usage in Analysis plane (f5.com) but this is could be still related to not enough resources for LTM + AWAF.

  • PhatANhappy's avatar
    PhatANhappy
    Icon for MVP rankMVP
    Jan 11, 2024

    ..... I would stay out of the gui as it tends to run hot and take more resources than its worth.
    that is only a dual core unit with a spinning disk drive,  so there is not a lot of hardware to work with.  Its also EOS - with 3 years left of life, meaning its a migration effort that will need to be repeated in the near future.

    from the CLI,  - top would be a better tool to work with.   

    https://my.f5.com/manage/s/article/K32264816

    what services are at the top few consuming all the resources?
    what is the "load " in top 1 min -5 min 
    what resourse is consuming the load?
    what is the uptime of the device?

    how many VIPS / objects do you have in your config?
    have you turned up logging?  are you useing HSL?

    what modules do you have provisioned?

    under provisioning - what is the size of the managment module? 

    if you are attempting to run ASM/AWAF - on the box - that is likely the cause, its far to small hardware wise to run efficently - you might find some relief here:   https://community.f5.com/t5/technical-forum/high-cpu-usage-independent-from-traffic/td-p/317138
     

    https://community.f5.com/t5/technical-articles/troubleshooting-high-cpu-utilisation-on-big-ip-systems/ta-p/280806

  • unavailable's avatar
    unavailable
    Icon for Cirrus rankCirrus
    Jan 11, 2024

    Hi PhatANhappy,

    many thanks for your fast reply.
     At the momemt there are 115 virtual servers configured on load balancer, there are 2 modules provisioned Ltm/Asm  both have been configuring with "nominal" provisioning.
    The hardware type is i2600, may be the problem could be refer to

    Bug ID 1060393: Extended high CPU usage caused by JavaScript Obfuscator  ?

    Rgds,

     

    • PhatANhappy's avatar
      PhatANhappy
      Icon for MVP rankMVP
      Jan 11, 2024

      that could be the issue.   definatly something to look at.
      i would lean more towards using the top command to find out what is causing the issue.
      With ASM -  i would further be concerned about spinning disks vs. SSD which will show up as a  'wa' in the top stastics.

    • Nikoolayy1's avatar
      Nikoolayy1
      Icon for MVP rankMVP
      Jan 11, 2024

      PhatANhappy  mentioned about too small resources and that this is control plane utlization not data plane and the ''top'' will review if it is the bug or another process. About the ''wa'' that is something I would also renember from this discussion.

       

      About the Analysis plane this in most cases is the ASM/AWAF policy builder learning, so also better also switch it off The BIG-IP ASM device shows high CPU usage in Analysis plane (f5.com) but this is could be still related to not enough resources for LTM + AWAF.

    • Nikoolayy1's avatar
      Nikoolayy1
      Icon for MVP rankMVP
      Jan 16, 2024

      Keep in mind if it is resource issue then the upgrade because of the reboot can clear it out for some time, so monitor the system for some time.