Automatic merge from submit-queue
add optional mutation checks for shared informer cache
We need to make sure that no one is mutating caches if they're using a shared informer. It is important that whatever is tracking those changes gets the object *before* anyone else possibly could.
This adds the ability to track the original objects in the cache and their current values. Go doesn't have an exit hook or a way to say "wait for non-daemon go-funcs to complete before exit", so this runs a gofunc on a loop that can panic the entire process. It's gated behind an env var.
@derekwaynecarr did I get the right spots to make sure that e2e runs with this flag?
@smarterclayton @kubernetes/rh-cluster-infra
Automatic merge from submit-queue
DeltaFIFO should not report HasSynced until deletes are reported
Deletions identified by knownObjects on the initial sync replace
(immediately after the List from a reflector) should be considered part
of the initial sync since we have the information available at the time
we do the deletion.
An error during Replace() can result in Populated not being set, but it
was incorrect before (population would be wrong) and queueActionLocked
does not error except on "my cache is broken because I gave an incorrect
keyFunc".
@lavalamp @deads2k hit this while trying to use DeltaFIFO to implement an "external" controller (one that uses the knownObjects as provided by a call to a remote system of record).
Not 1.4
Deletions identified by knownObjects on the initial sync replace
(immediately after the List from a reflector) should be considered part
of the initial sync since we have the information available at the time
we do the deletion.
An error during Replace() can result in Populated not being set, but it
was incorrect before (population would be wrong) and queueActionLocked
does not error except on "my cache is broken because I gave an incorrect
keyFunc".
1. When overlapping deployments are discovered, annotate them
2. Expose those overlapping annotations as warnings in kubectl describe
3. Only respect the earliest updated one (skip syncing all other overlapping deployments)
4. Use indexer instead of store for deployment lister
Automatic merge from submit-queue
Do not hold the lock for a long time
Followup to #30839.
I'm not convinced this is a super great idea but I'll throw it out and let others decide.
Ref https://github.com/kubernetes/minikube/issues/368
Ref #30759
Automatic merge from submit-queue
Change storeToNodeConditionLister to return []*api.Node instead of api.NodeList for performance
Currently copies that are made while copying/creating api.NodeList are significant part of scheduler profile, and a bunch of them are made in places, that are not-parallelizable.
Ref #28590
Automatic merge from submit-queue
Some scheduler optimizations
Ref #28590
This PR doesn't do anything fancy - it is just reducing amount of memory allocations in scheduler, which in turn significantly speeds up scheduler.
Automatic merge from submit-queue
optimize deleteFromIndices method of thread_safe_store
As all methods of thread_safe_store are threadsafe, so i think, in deleteFromIndices method, if the index is nil, need not run the for structure below
When kubelet starts a pod that refers to non-existing PV, PVC or Node, it
should clearly show that the requested element does not exist.
Previous "PersistentVolumeClaim 'default/ceph-claim-wm' is not in cache"
looks like random kubelet hiccup, while "PersistentVolumeClaim
'default/ceph-claim-wm' not found" suggests that the object may not exist at
all and it might be an user error.
Fixes#27523
Refactor storePodsNamespacer.List() and
storeReplicationContollersNamespacer.List(). They are the same
function, just with different signatures.
This fixes a bug where, when we fell back on a brute force approach, we
were still returning an error.
Also change to explicit return without named return values.
A recent change tries to separate resync and relist. The motivation
was to avoid triggering relist when a resync is required.
However, the change is not effective since it stops the watcher. As hongchao
mentioned in the original comment, today's storage interface will not deliever
any progress notification to the watch chan. So any watcher that does not receive
events for the last few seconds will not be able to catch up from the previous
index after a hard close since the index of the last received event is out of
the cache window inside etcd2.
This pull request tries to fix this issue by not stoping watcher when a resync is
required.
Automatic merge from submit-queue
add namespace index for cache
@wojtek-t
Implement in this approach make the change of lister.go small, but we should replace all `NewInformer()` to `NewIndexInformer()`, even when someone not want to filter by namespace(eg. gc_controller and scheduler). Any suggestion?
Reflectors started from goroutines are broken because Go doesn't allow
runtime.Callers to see the spawning goroutine. Do a best effort parse of
the call stack for now.
Add logging so that we can easily see which reflectors processes launch,
and measure in logs the frequency of sync intervals.
The message as it is framed right now does not make any sense for the
end users of our system. It might even lead to confusion. So this is
attempt to make the error message less confusing.
Update the Deployments' API types, defaulting code, conversions, helpers
and validation to use ReplicaSets instead of ReplicationControllers and
LabelSelector instead of map[string]string for selectors.
Also update the Deployment controller, registry, kubectl subcommands,
client listers package and e2e tests to use ReplicaSets and
LabelSelector for Deployments.
Move type LabelSelector and type LabelSelectorRequirement from pkg/apis/extensions
This avoids an import loop when Job (and later DaemonSet, Deployment, ReplicaSet)
are moved out of extensions to new api groups.
Also Move LabelSelectorAsSelector utility from pkg/apis/extensions/ to pkg/api/unversioned/
Also its test.
Also LabelSelectorOp* constants.
Also the pkg/apis/extensions/validation functions ValidateLabelSelectorRequirement and
ValidateLabelSelector move to pkg/api/unversioned
The related type in pkg/apis/extensions/v1beta1/ is staying there. I might move
it in another PR if neccessary.
