-
Bug
-
Resolution: Unresolved
-
Normal
-
None
-
4.14.z
Description of problem:
When I deploy a QoS Guaranteed with cpu-c-states.crio.io: disabled pod annotation, and when there has been a container restart e.g. in the describe the container failed once ``` Normal Pulled 2m54s kubelet Successfully pulled image "quay.io/karampok/dpdk:latest" in 862.099864ms (862.110666ms including waiting) Normal Created 2m53s (x2 over 2m55s) kubelet Created container io Normal Started 2m53s (x2 over 2m55s) kubelet Started container io ``` the value of pm_qos_resume_latency_us is not restored after a delete ``` cat /var/run/crio/cpu/cpu10/power/pm_qos_resume_latency_us n/a #while the default should be 0 ``` Same for the `cpu-freq-governor.crio.io:` but that less impact as it not keep the core in C0 when idle
Version-Release number of selected component (if applicable):
How reproducible:
Deploy a guaranteed pods that use these annotations and ave a container restart.
Steps to Reproduce:
1. 2. 3.
Actual results:
Expected results:
Additional info: