Describe the solution you'd like
with following config:
policy:
Active: balloons
AvailableResources:
CPU: cpuset:1-63
ReservedResources:
CPU: cpuset:63
after I scheduled new pod I got an error
unable to apply cgroup configuration: failed to write "63": write /sys/fs/cgroup/cpuset/docker/6eb0031417c8e4e3fd56ba13702cd6cdb5bd920d6065d79a2ca7afe8ca2b40d0/kubelet.slice/kubelet-kubepods.slice/kubelet-ku
because my machine has only: 56 cpus
Why this is needed
I want to prevent situation when early detected misconfiguration can be validate before we reconfigure kubelet or even try to run.
ps. I don't know is this generic issues I found it on this branch: may be related to "new feature" cputree developed there: https://github.com/askervin/cri-resource-manager/tree/5Lp_WIP_balloons_topology_fix
Describe the solution you'd like
with following config:
after I scheduled new pod I got an error
because my machine has only: 56 cpus
Why this is needed
I want to prevent situation when early detected misconfiguration can be validate before we reconfigure kubelet or even try to run.
ps. I don't know is this generic issues I found it on this branch: may be related to "new feature" cputree developed there: https://github.com/askervin/cri-resource-manager/tree/5Lp_WIP_balloons_topology_fix