Introduce new flag "--include-uninitialized" to kubectl by dixudx · Pull Request #50497 · kubernetes/kubernetes

dixudx · 2017-08-11T06:28:06Z

What this PR does / why we need it:

Introduce --include-uninitialized as a global flag to kubectl

Which issue this PR fixes (optional, in fixes #<issue number>(, fixes #<issue_number>, ...) format, will close that issue when PR gets merged): fixes #49035

Special notes for your reviewer:
/assign @caesarxuchao @smarterclayton @ahmetb @deads2k

Release note:

Add flag "--include-uninitialized" to kubectl annotate, apply, edit-last-applied, delete, describe, edit, get, label, set. "--include-uninitialized=true" makes kubectl commands apply to uninitialized objects, which by default are ignored if the names of the objects are not provided. "--all" also makes kubectl commands apply to uninitialized objects. Please see the [initializer](https://kubernetes.io/docs/admin/extensible-admission-controllers/) doc for more details.

k8s-ci-robot · 2017-08-11T06:28:13Z

Hi @dixudx. Thanks for your PR.

I'm waiting for a kubernetes member to verify that this patch is reasonable to test. If it is, they should reply with /ok-to-test on its own line. Until that is done, I will not automatically test new commits in this PR, but the usual testing commands by org members will still work. Regular contributors should join the org to skip this step.

I understand the commands that are listed here.

Details

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. I understand the commands that are listed here.

ahmetb · 2017-08-14T21:54:41Z

 		NamespaceParam(cmdNamespace).DefaultNamespace().AllNamespaces(allNamespaces).
 		FilenameParam(enforceNamespace, options).
 		SelectorParam(selector).
+		IncludeUninitializedParam(includeUninitialized).


I think the proposal #49035 (comment) intends to disable --include-uninitialized for batch commands. For example:

kubectl describe pods/foo: must include uninitialized

kubectl describe pods: must NOT include uninitialized because it lists pods.

@ahmetb As @caesarxuchao commented in #49035

Both should includes uninitialized objects. --include-uninitialized doesn't affect them at all.

So I keep it unchanged.

If user explicitly set --include-uninitialized=false when kubectl describe pods, then we should skip uninitialized pods.

ahmetb · 2017-08-14T21:57:10Z

 		Example: getExample,
 		Run: func(cmd *cobra.Command, args []string) {
+			includeUninitialized := cmdutil.GetFlagBool(cmd, "include-uninitialized")
+			fmt.Printf("L122 %+v\n", includeUninitialized)


forgot to remove this?

ahmetb · 2017-08-14T22:03:07Z

 		return err
 	}

+	includeUninitialized := cmdutil.GetFlagBool(cmd, "include-uninitialized")


See my comment at #49035 (comment) I think annotate works with individual resources (and doesn't work in batch mode) so it should not have this flag.

ahmetb · 2017-08-14T22:03:13Z

 		return err
 	}

+	includeUninitialized := cmdutil.GetFlagBool(cmd, "include-uninitialized")


See my comment at #49035 (comment) I think label works with individual resources (and doesn't work in batch mode) so it should not have this flag.

ahmetb · 2017-08-14T22:03:17Z

 		return err
 	}

+	includeUninitialized := cmdutil.GetFlagBool(cmd, "include-uninitialized")


See my comment at #49035 (comment) I think set works with individual resources (and doesn't work in batch mode) so it should not have this flag.

ditto for other set_*.go changes.

ahmetb · 2017-08-14T22:04:17Z


 	export bool

+	includeUninitialized bool


I think you can group this next to selectAll above.

ahmetb · 2017-08-14T22:10:15Z

-		Timeout:        FlagInfo{prefix + FlagTimeout, "", "0", "The length of time to wait before giving up on a single server request. Non-zero values should contain a corresponding time unit (e.g. 1s, 2m, 3h). A value of zero means don't timeout requests."},
+		CurrentContext:       FlagInfo{prefix + FlagContext, "", "", "The name of the kubeconfig context to use"},
+		Timeout:              FlagInfo{prefix + FlagTimeout, "", "0", "The length of time to wait before giving up on a single server request. Non-zero values should contain a corresponding time unit (e.g. 1s, 2m, 3h). A value of zero means don't timeout requests."},
+		IncludeUninitialized: FlagInfo{prefix + FlagIncludeUninitialized, "", "false", "If true, will include the objects which have a non-empty initializer list"},


If true, will include the objects which have a non-empty initializer list

I think this captures the implementation detail, and not necessarily user-friendly.

Also:

will word doesn't match the rest of the flags documentation

object should be object(s)

sentence should be completed with a period (e.g. see kubectl get --help)

How about:

If true, include the object(s) that have not been realized yet. Such object(s) have a non-empty initializer list.

Also question:

I see the default value is false. Does this mean kubectl get pods --include-uninitialized still means includeUnintialized=false? Should the user explicitly set =true or does this work like --all-namespaces already with this code?

@ahmetb

Does this mean kubectl get pods --include-uninitialized still means includeUnintialized=false?

No, they are totally different. kubectl get pods --include-uninitialized means includeUnintialized is set to True.

kubectl get pods --include-uninitialized is identical to kubectl get pods --include-uninitialized=true.

does this work like --all-namespaces already with this code

Yes, you're right. This works like other cobra-based CLIs, including our kubectl commands, like you mentioned --all-namespaces.

dixudx · 2017-08-15T07:02:42Z

Updated. PTAL.

caesarxuchao · 2017-08-15T16:49:55Z

I'll take a look today.

caesarxuchao

Thanks @dixudx. We also need to add either test-cmd.sh or e2e tests. @mengqiy do you know where we should add tests?

@mengqiy could you also take a look at this PR or find someone from sig-cli to review it? I'm not that familiar with kubectl code. Thanks.

caesarxuchao · 2017-08-16T00:28:56Z

+		// unless explicitly set --include-uninitialized=false
+		includeUninitialized = true
+	}
+	if o.selector != "" {


When both specified, does o.selector overwrite o.all?

@caesarxuchao Yes, this is current implementation.

But we have to discuss on such combinations. I think o.all is the dominant one. When other flags combine with o.all, includeUninitialized should be set to true.

ping @ahmetb @smarterclayton @pwittrock, PTAL.

I think it's strange if kubectl annotate pod -l foo=bar returns different results from kubectl annotate pod -l foo=bar --all. I think the current implementation is good.

caesarxuchao · 2017-08-16T00:33:03Z

 			p.backOff.Sleep(backOffPeriod)
 		}
-		current, getErr = p.helper.Get(namespace, name, false)
+		current, getErr = p.helper.Get(namespace, name, false, false)


If name is specified, we should always return the object to the user, no matter if it's initialized. We shouldn't change Get().

caesarxuchao · 2017-08-16T00:45:50Z

 		NamespaceParam(cmdNamespace).DefaultNamespace().AllNamespaces(allNamespaces).
 		FilenameParam(enforceNamespace, options).
 		SelectorParam(selector).
+		IncludeUninitializedParam(includeUninitialized).


If user explicitly set --include-uninitialized=false when kubectl describe pods, then we should skip uninitialized pods.

caesarxuchao · 2017-08-16T00:57:55Z

 	selector := cmdutil.GetFlagString(cmd, "selector")
 	allNamespaces := cmdutil.GetFlagBool(cmd, "all-namespaces")
 	showKind := cmdutil.GetFlagBool(cmd, "show-kind")
+	showAll := cmdutil.GetFlagBool(cmd, "show-all")


I just noticed --show-all exists in many other commands. I don't know if it's alright we only special case kubectl get <resource> --show-all.

On the other hand, kubectl get is quite special, it doesn't accept --all.

It seems we can't update the help text of --show-all for only kubectl get.

@mengqiy @pwittrock do you know if --show-all is supposed to only affect pods? The helper text reads "When printing, show all resources (default hide terminated pods.)".

Discussed with @pwittrock, let's don't specialize --show-all for kubectl get. That is, kubectl get pods --show-all doesn't include uninitialized objects. Users have to use kubectl get pods --include-uninitialized=true.

The root problem is that kubectl get doesn't have a --all flag, while many other kubectl commands have. We need to unify the flags, but that's another PR.

@caesarxuchao What about selector? Seems we have not taken selector into consideration.

For selector, we should make it consistent with kubectl label. That is "Does not include the uninitialized objects by default, unless user explicitly set --include-uninitialized=true". Any comments?

caesarxuchao · 2017-08-16T01:02:23Z

 }

-func (m *Helper) Get(namespace, name string, export bool) (runtime.Object, error) {
+func (m *Helper) Get(namespace, name string, export, includeUninitialized bool) (runtime.Object, error) {


Let's leave this function as is. I don't see any use case to hide an uninitialized object if user specifically asks for it.

caesarxuchao · 2017-08-16T01:03:59Z

+		Namespace:            namespace,
+		Name:                 name,
+		Export:               export,
+		IncludeUninitialized: includeUninitialized,


Same here. If name is specified, we always show user the object.

caesarxuchao · 2017-08-16T01:11:43Z

+	Context              clientcmdapi.Context
+	CurrentContext       string
+	Timeout              string
+	IncludeUninitialized bool


The config here is not specific for kubectl, so we shouldn't change it. Probably you should add it in helpers.go like other flags.

caesarxuchao · 2017-08-16T01:16:12Z

 	}

+	var includeUninitialized bool
+	if o.all {


We need to update the flag comment of --all.

ahmetb · 2017-08-16T17:12:23Z

+		// unless explicitly set --include-uninitialized=false
+		includeUninitialized = true
+	}
+	if o.Selector != "" {


is Selector != "" when I do kubectl set pods/foo ...? Trying to understand what this clause is for.

I think so, "foo" will be a name, not a selector in this case.

mengqiy · 2017-08-16T17:56:43Z

@dixudx @caesarxuchao I will review it today.
/assign
/ok-to-test

smarterclayton · 2017-08-17T02:16:59Z

 		NamespaceParam(cmdNamespace).DefaultNamespace().
 		FilenameParam(enforceNamespace, &options.FilenameOptions).
 		SelectorParam(options.Selector).
+		IncludeUninitializedParam(includeUninitialized).


This is not a param if you're doing parsing yourself. It would just be IncludeUninitialized.

smarterclayton · 2017-08-17T02:17:37Z

 		return err
 	}

+	var includeUninitialized bool


This is a lot of duplicated code that is effectively the same. It needs to be unified - copying and pasting this code into many places (with some differences) is a maintenance problem.

dixudx · 2017-09-01T07:03:23Z

Add another testcase for apply --prune --include-uninitialized --all.

@caesarxuchao @mengqiy @pwittrock PTAL. Thanks.

ahmetb · 2017-09-01T07:45:06Z

@dixudx consider adding spaces after the commas in the release note.

mengqiy · 2017-09-01T07:46:51Z

+  kube::log::status "Testing --include-uninitialized"
+
+  ### Create a deployment
+  kubectl create --request-timeout=1 -f hack/testdata/initializer-deployments.yaml 2>&1 "${kube_flags[@]}" || true


From help text:

--request-timeout='0': The length of time to wait before giving up on a single server request. Non-zero values should contain a corresponding time unit (e.g. 1s, 2m, 3h). A value of zero means don't timeout requests.

I don't understand why you give --request-timeout=1. And it's not clear to me that timeout is 1s or 1m or ...

I think it's b/c I left a comment about it. If we don't specify a timeout, the test spend 30s here for the blocked initializer. Now I'm suspecting this 1s might cause flakes, I'm not sure making it another value like 5s would help prevent the flake at all.

I don't understand why you give --request-timeout=1

@mengqiy I've added pending initializers in the yaml, which will keep blocking until timeout. Per discussion.

And it's not clear to me that timeout is 1s or 1m or ...

Just follow other commands in the test files. Actually --request-timeout=1 means 1sec.

Since the flag description of --request-timeout is quite confusing. We may consider to elaborate it.

Make sense.

dixudx · 2017-09-01T08:11:11Z

/test pull-kubernetes-e2e-gce-bazel

mengqiy · 2017-09-01T16:18:39Z

@dixudx Please squash some unnecessary commits.

Don't block on this, it's code freeze today.

caesarxuchao · 2017-09-01T17:50:44Z

RE. release note, I think we need to provide more information. ~~How about:~~
[edit] I updated the release note, since today is code freeze.

Add flag "--include-uninitialized" to kubectl annotate, apply, edit-last-applied, delete, describe, edit, get, label, set. "--include-uninitialized=true" makes kubectl commands apply to uninitialized objects, which by default are ignored if the names of the objects are not provided. "--all" also makes kubectl commands apply to uninitialized objects. Please see the [initializer](https://kubernetes.io/docs/admin/extensible-admission-controllers/) doc for more details.

We need to update the linked doc later.

caesarxuchao · 2017-09-01T17:52:46Z

@deads2k could you approve? Thanks.

(We had Clayton and Phil approved the flags in the issue #49035)

mengqiy · 2017-09-01T17:57:25Z

/lgtm

deads2k · 2017-09-01T18:23:06Z

/approve

k8s-github-robot · 2017-09-01T18:34:05Z

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: caesarxuchao, deads2k, dixudx, mengqiy

Associated issue: 49035

The full list of commands accepted by this bot can be found here.

Details

Needs approval from an approver in each of these OWNERS Files:

~~hack/OWNERS~~ [deads2k]
~~pkg/kubectl/OWNERS~~ [deads2k]
~~staging/src/k8s.io/client-go/OWNERS~~ [caesarxuchao,deads2k]

You can indicate your approval by writing /approve in a comment
You can cancel your approval by writing /approve cancel in a comment

k8s-github-robot · 2017-09-03T06:39:10Z

/test all [submit-queue is verifying that this PR is safe to merge]

k8s-github-robot · 2017-09-03T06:49:59Z

Automatic merge from submit-queue (batch tested with PRs 51301, 50497, 50112, 48184, 50993)

k8s-ci-robot · 2017-09-03T07:58:20Z

@dixudx: The following test failed, say /retest to rerun them all:

Test name	Commit	Details	Rerun command
pull-kubernetes-e2e-kops-aws	`d80ff0f`	link	`/test pull-kubernetes-e2e-kops-aws`

Full PR test history. Your PR dashboard. Please help us cut down on flakes by linking to an open issue when you hit one in your PR.

Details

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. I understand the commands that are listed here.

@caesarxuchao

Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>. kubectl should return an error if "-l" and "--all" are both specified **What this PR does / why we need it**: Per discussion in [kubernetes#50497](kubernetes#50497 (comment)) **Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes # **Special notes for your reviewer**: /assign @caesarxuchao @mengqiy **Release note**: ```release-note kubectl should return an error if "-l" and "--all" are both specified ```

@caesarxuchao

…tialized Automatic merge from submit-queue (batch tested with PRs 51301, 50497, 50112, 48184, 50993) Introduce new flag "--include-uninitialized" to kubectl **What this PR does / why we need it**: Introduce `--include-uninitialized` as a global flag to kubectl **Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes kubernetes#49035 **Special notes for your reviewer**: /assign @caesarxuchao @smarterclayton @ahmetb @deads2k **Release note**: ```release-note Add flag "--include-uninitialized" to kubectl annotate, apply, edit-last-applied, delete, describe, edit, get, label, set. "--include-uninitialized=true" makes kubectl commands apply to uninitialized objects, which by default are ignored if the names of the objects are not provided. "--all" also makes kubectl commands apply to uninitialized objects. Please see the [initializer](https://kubernetes.io/docs/admin/extensible-admission-controllers/) doc for more details. ```

@caesarxuchao

Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>. kubectl should return an error if "-l" and "--all" are both specified **What this PR does / why we need it**: Per discussion in [kubernetes#50497](kubernetes#50497 (comment)) **Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes # **Special notes for your reviewer**: /assign @caesarxuchao @mengqiy **Release note**: ```release-note kubectl should return an error if "-l" and "--all" are both specified ```

k8s-ci-robot assigned ahmetb, caesarxuchao, deads2k and smarterclayton Aug 11, 2017

k8s-ci-robot added the cncf-cla: yes Indicates the PR's author has signed the CNCF CLA. label Aug 11, 2017

k8s-ci-robot added the needs-ok-to-test Indicates a PR that requires an org member to verify it is safe to test. label Aug 11, 2017

k8s-github-robot added size/L Denotes a PR that changes 100-499 lines, ignoring generated files. release-note Denotes a PR that will be considered when it comes time to generate release notes. labels Aug 11, 2017

dixudx mentioned this pull request Aug 11, 2017

kubectl delete --all: does not include uninitialized resources #49035

Closed

dixudx force-pushed the kubectl-include-uninitialized branch from 4698d59 to 9cda6b1 Compare August 12, 2017 01:57

dixudx changed the title ~~[WIP] Kubectl include uninitialized~~ Kubectl include uninitialized Aug 12, 2017

dixudx changed the title ~~Kubectl include uninitialized~~ Introduce --include-uninitialized as a global flag to kubectl Aug 12, 2017

ahmetb suggested changes Aug 14, 2017

View reviewed changes

dixudx force-pushed the kubectl-include-uninitialized branch from 9cda6b1 to b948b85 Compare August 15, 2017 06:40

caesarxuchao reviewed Aug 16, 2017

View reviewed changes

dixudx force-pushed the kubectl-include-uninitialized branch from b948b85 to 562f099 Compare August 16, 2017 14:31

k8s-github-robot added the needs-rebase Indicates a PR cannot be merged because it has merge conflicts with HEAD. label Aug 16, 2017

dixudx force-pushed the kubectl-include-uninitialized branch from 562f099 to 4c1c366 Compare August 16, 2017 14:40

k8s-github-robot removed the needs-rebase Indicates a PR cannot be merged because it has merge conflicts with HEAD. label Aug 16, 2017

ahmetb reviewed Aug 16, 2017

View reviewed changes

k8s-ci-robot removed the needs-ok-to-test Indicates a PR that requires an org member to verify it is safe to test. label Aug 16, 2017

k8s-ci-robot assigned mengqiy Aug 16, 2017

dixudx force-pushed the kubectl-include-uninitialized branch from 4c1c366 to a4f0b47 Compare August 17, 2017 01:55

dixudx mentioned this pull request Aug 17, 2017

Kubectl get show uninitialized resources #50445

Closed

smarterclayton reviewed Aug 17, 2017

View reviewed changes

dixudx changed the title ~~Introduce --include-uninitialized as a global flag to kubectl~~ Introduce new flag "--include-uninitialized" to kubectl Sep 1, 2017

dixudx force-pushed the kubectl-include-uninitialized branch from 93feafc to eab8c1d Compare September 1, 2017 07:01

k8s-github-robot removed the lgtm "Looks good to me", indicates that a PR is ready to be merged. label Sep 1, 2017

test fix

d80ff0f

dixudx force-pushed the kubectl-include-uninitialized branch from eab8c1d to d80ff0f Compare September 1, 2017 07:05

mengqiy reviewed Sep 1, 2017

View reviewed changes

k8s-ci-robot added the lgtm "Looks good to me", indicates that a PR is ready to be merged. label Sep 1, 2017

k8s-github-robot added the approved Indicates a PR has been approved by an approver from all required OWNERS files. label Sep 1, 2017

abgworrall modified the milestone: v1.8 Sep 2, 2017

k8s-github-robot merged commit 5c0b265 into kubernetes:master Sep 3, 2017

dixudx deleted the kubectl-include-uninitialized branch September 3, 2017 06:52

dixudx mentioned this pull request Sep 4, 2017

InitializerConfiguration will impact deployment #51883

Closed

gyliu513 mentioned this pull request Sep 4, 2017

kubectl get: does not show uninitialized resources #49032

Closed

dixudx mentioned this pull request Jul 26, 2021

cleanup description on deprecated include-uninitialized flag #103905

Merged

Uh oh!

Conversation

dixudx commented Aug 11, 2017 • edited by caesarxuchao Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

k8s-ci-robot commented Aug 11, 2017

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

dixudx commented Aug 15, 2017

Uh oh!

caesarxuchao commented Aug 15, 2017

Uh oh!

caesarxuchao left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

mengqiy commented Aug 16, 2017

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

dixudx commented Sep 1, 2017

Uh oh!

ahmetb commented Sep 1, 2017

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ahmetb Sep 1, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

dixudx commented Aug 11, 2017 •

edited by caesarxuchao

Loading

ahmetb Sep 1, 2017 •

edited

Loading

mengqiy commented Sep 1, 2017 •

edited

Loading

caesarxuchao commented Sep 1, 2017 •

edited

Loading

caesarxuchao commented Sep 1, 2017 •

edited

Loading