This improves daemon.log retention from a few hours to week
(i.e. without log-rotation). This enables debugability support
if customer stays on Kubernetes 1.24 or 1.25 for any period of
time.
This is a manual backport of PRs 110496 and 117241 by Kir Kolyshkin.
Cherry-pick of: https://github.com/kubernetes/kubernetes/pull/117892/
([1.24] vendor: bump runc to 1.1.6)
Cherry-pick of: https://github.com/kubernetes/kubernetes/pull/117682/
([1.25] vendor: bump runc to 1.1.6)
This change is not required in Kubernetes 1.26 upwards.
This issue comes from the integration of the Misc controller in
Kernel 5.13 . Problem comes from a discrepancy between the code
creating the "Misc" Cgroup and the code cleaning the Cgroup that
doesn't handle the "Misc" Cgroup, leaving it behind. runc integrated
a fix in 1.1.6, HOWEVER kubelet depends on runc's cgroup libraries.
In order to clean up pods using the new "Misc" controller,
runc cgroup library need to be updated to be aware of it.
So even if our system run 1.1.6 but kubelet is not build
with these library, the problem occurs.
See more details at:
https://www.suse.com/support/kb/doc/?id=000021270https://github.com/opencontainers/runc/issues/3849#issuecomment-1544519250
Test Plan
PASS: Install iso as AIO-SX
PASS: Build-pkg 1.24 and 1.25 successfully.
PASS: AIO-SX USM upgrade with K8S 1.24. Verify no log flood.
PASS: AIO-SX: K8S Upgrade from 1.24 to 1.25, verify no log flood.
PASS: make WHAT="cmd/kube-proxy cmd/kube-apiserver \
cmd/kube-controller-manager cmd/kubelet cmd/kubeadm \
cmd/kube-scheduler cmd/kubectl cmd/genman"
Closes-Bug: 2098478
Change-Id: Iea84aceef74cad45f514bb987010e12880645582
Signed-off-by: Saba Touheed Mujawar <sabatouheed.mujawar@windriver.com>
6f490796eb