Last Modified: Jun 10, 2021
See more info
Known Affected Versions:
13.1.0, 126.96.36.199, 188.8.131.52, 184.108.40.206, 220.127.116.11, 18.104.22.168, 22.214.171.124, 126.96.36.199, 188.8.131.52, 13.1.1, 184.108.40.206, 220.127.116.11, 18.104.22.168, 22.214.171.124, 13.1.3, 126.96.36.199, 188.8.131.52, 184.108.40.206, 220.127.116.11, 18.104.22.168, 22.214.171.124, 13.1.4, 126.96.36.199
Opened: Jun 03, 2019
The TMSTAT CPU usage values are not always in agreement with the output from 'top'.
When HTSplit is active, the even-numbered CPUs are data plane CPUs and the odd-numbered CPUs are control plane, and the highest CPU is reserved for analytics. The CPU usage is reported as the highest of either the data or the control plane (the analytics CPU is not included in calculations) because when a system is busy you are out of 'head room' if either the data or the control plane is too busy. In a heavily loaded system (with or without HTSPlit), the statistics calculations may not complete in one cycle time, so the CPU values might lag behind what 'top' reports. This results in TMSTAT CPU calculations that are not the same as the values displayed in 'top'.
This can occur on: -- BIG-IP hardware platform systems that use HTSplit. -- Heavily loaded systems when the control plane is not keeping the calculations in sync with the output of 'top'. -- BIG-IP Virtual Edition (VE) with significantly uneven traffic handling across all TMMs/CPU Cores.