The HPA controller keeps a flat history of recommendations for
stabilization. However when both up and down scale stabilization are
configured, the interpretation of the history changes depending on the
direction of movement. What we want is to keep the stabilized
recommendation within the envelope of the minimum and maximum over
configured stabilization windows. We should only move when the
envelope forces a move.
in pkg/scheduler/eventhandlers.go
* add event for unscheduled pod
* delete event for unscheduled pod
* add event for scheduled pod
* delete event for scheduled pod
in pkg/scheduler/framework/plugins/defaultbinder/default_binder.go
* Attempting to bind pod to node
in pkg/scheduler/scheduler.go
* Updating pod condition
* Attempting to schedule pod
* Skip schedule deleting pod
fakeVolumeHost previously implemented both the KubeletVolumeHost and
AttachDetachVolumeHost interfaces. This design makes it difficult to test the
CSIAttacher since it behaves differently depending on what type of
VolumeHost is supplied.
When the maxsurge daemonset gate is disabled, the registry and validation
must properly handle stripping the field. In the special case where that
would leave the MaxUnavailable field set to 0, we must set it to 1 which
is the default value.
The MaxSurge field on DaemonSet rolling updates allows a daemonset
workload to have two pods running simultaneously on a node during
an update in order to perform zero-disruption handoffs of client
traffic.
In case /var/lib/kubelet is a symlink, "losetup -j <device in
/var/lib/kubelet>" will show the device paths with symlinks fully
evaluated.
Fix the parsing routine and expand symlinks in the path that we want to
find in "losetup -j" output.
The pInfo.Timestamp is refreshed but the sort in activeQ or podBackoffQ is not be updated when pod is already present in the backoff or active queue.
AddUnschedulableIfNotPresent() return error if pod is already present in the backoff or active queue, and there is no re-add.
So refresh pInfo.Timestamp when the pod is not present in the three sub-queues, otherwise need to update the order of the pod in the active or backoff queue, for example p.activeQ.Update(pInfo)
Deprecated metrics are removed and suggest to use the Histogram
metrics got from scheduler extension points.
Signed-off-by: Dave Chen <dave.chen@arm.com>
Co-authored-by: wawa0210 <xiaozhang0210@hotmail.com>
