The "none" driver is expected to be used in "rootless" mode until OCI/CRI runtime
get support for cgroup2 (unified) mode with nsdelegate.
Even after cgroup2 gets supported in the ecosystem, the "none" driver will remain
because nested containers might not always get support for cgroup2 (via systemd).
Signed-off-by: Akihiro Suda <akihiro.suda.cz@hco.ntt.co.jp>
# Conflicts:
# cmd/kubelet/app/server.go
# pkg/features/kube_features.go
The metrics structure passed to volume stat calculator
can contain real stats on subset of metrics fields. For example,
the metrics structure filled by a CSI driver can have
either INODES or BYTES filled, IOW it a valid return.
In such cases the volume stat calculator panic with below
trace:
0516 21:36:19.013143 14452 runtime.go:69] Observed a panic: "invalid memory address or nil pointer dereference" (runtime error: invalid memory address or nil pointer dereference)
/home/hchiramm/gopath/src/k8s.io/kubernetes/_output/local/go/src/k8s.io/kubernetes/vendor/k8s.io/apimachinery/pkg/util/runtime/runtime.go:76
/home/hchiramm/gopath/src/k8s.io/kubernetes/_output/local/go/src/k8s.io/kubernetes/vendor/k8s.io/apimachinery/pkg/util/runtime/runtime.go:65
/home/hchiramm/gopath/src/k8s.io/kubernetes/_output/local/go/src/k8s.io/kubernetes/vendor/k8s.io/apimachinery/pkg/util/runtime/runtime.go:51
/usr/local/go/src/runtime/panic.go:522
/usr/local/go/src/runtime/panic.go:82
/usr/local/go/src/runtime/signal_unix.go:390
/home/hchiramm/gopath/src/k8s.io/kubernetes/_output/local/go/src/k8s.io/kubernetes/vendor/k8s.io/apimachinery/pkg/api/resource/quantity.go:697
/home/hchiramm/gopath/src/k8s.io/kubernetes/_output/local/go/src/k8s.io/kubernetes/vendor/k8s.io/apimachinery/pkg/api/resource/quantity.go:685
/home/hchiramm/gopath/src/k8s.io/kubernetes/_output/local/go/src/k8s.io/kubernetes/pkg/kubelet/server/stats/volume_stat_calculator.go:144
/home/hchiramm/gopath/src/k8s.io/kubernetes/_output/local/go/src/k8s.io/kubernetes/pkg/kubelet/server/stats/volume_stat_calculator.go:125
/home/hchiramm/gopath/src/k8s.io/kubernetes/_output/local/go/src/k8s.io/kubernetes/pkg/kubelet/server/stats/volume_stat_calculator.go:65
/home/hchiramm/gopath/src/k8s.io/kubernetes/_output/local/go/src/k8s.io/kubernetes/vendor/k8s.io/apimachinery/pkg/util/wait/wait.go:152
/home/hchiramm/gopath/src/k8s.io/kubernetes/_output/local/go/src/k8s.io/kubernetes/vendor/k8s.io/apimachinery/pkg/util/wait/wait.go:153
/usr/local/go/src/runtime/asm_amd64.s:1337
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
If kubelet never gets past sandbox creation (i.e., never attempted to
create containers for a pod), it should retry the sandbox creation on
failure, regardless of the restart policy of the pod.
On Windows the only way to access the container's network interfaces is
by running another process in the pod from which we can use a
netcat-like program to proxy the TCP stream
Proposed wincat.exe can be found here: https://github.com/benmoss/wincat
These updates are based on discussions had about the preferred semantics
of the TopologyManager and will be reflected in changes to an upcoming
PR that adds the actual TopologyManager implementation.
Darren Shepherd
Erik Wilson
Akihiro Suda
Kubernetes Prow Robot
wojtekt
Ted Yu
Angela Li
Humble Chirammal
Rong Gao
Odin Ugedal
Yu-Ju Hong
rafatio
David Ashpole
Hemant Kumar
Ben Moss
Tara Gu
Renaud Gaubert
Kevin Klues
nolancon