[SPARK-33012][BUILD][K8S] Upgrade fabric8 to 4.10.3 #29888

laflechejonathan · 2020-09-28T07:31:46Z

What changes were proposed in this pull request?

This PR aims to upgrade kubernetes-client library to track fabric8's declared compatibility for k8s 1.18.0:
https://github.com/fabric8io/kubernetes-client#compatibility-matrix

Why are the changes needed?

According to fabric8, 4.9.2 is incompatible with k8s 1.18.0.

Does this PR introduce any user-facing change?

No.

How was this patch tested?

Not tested yet.

srowen · 2020-09-28T13:38:10Z

resource-managers/kubernetes/core/pom.xml

    </dependency>
    <!-- End of shaded deps. -->

+    <dependency>


Why does this become necessary, out of curiosity? looks like a downgrade. I'm just wondering if it affects anything else that depends on okhttp.

I agree with @srowen . This looks suspicious.

this was my mistake - reverted.

srowen · 2020-09-28T13:39:12Z

dev/deps/spark-deps-hadoop-3.2-hive-2.3

+kubernetes-model-admissionregistration/4.10.3//kubernetes-model-admissionregistration-4.10.3.jar
+kubernetes-model-apiextensions/4.10.3//kubernetes-model-apiextensions-4.10.3.jar
+kubernetes-model-apps/4.10.3//kubernetes-model-apps-4.10.3.jar
+kubernetes-model-autoscaling/4.10.3//kubernetes-model-autoscaling-4.10.3.jar


Are all of these additional deps needed - if they really aren't we might exclude them for tidiness.

I don't think we have a choice here, it looks like kubernetes-client split their jar into smaller dependencies:
fabric8io/kubernetes-client#2108, and since we depend on the top level kubernetes-client, we must pull these in.

dongjoon-hyun · 2020-09-28T15:52:25Z

ok to test

resource-managers/kubernetes/core/pom.xml

dongjoon-hyun

Thank you for your first contribution, @laflechejonathan .

Please don't remove the .
Which K8s server version are you using with this fabric8 library?
Although the matrix shows like that, Apache Spark master branch seems to work with K8s v1.18.8 without this upgrade, doesn't it?

SparkQA · 2020-09-28T16:47:22Z

Kubernetes integration test starting
URL: https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder-K8s/33806/

SparkQA · 2020-09-28T17:10:33Z

Kubernetes integration test status failure
URL: https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder-K8s/33806/

SparkQA · 2020-09-28T18:35:18Z

Test build #129191 has finished for PR 29888 at commit ee845e0.

This patch passes all tests.
This patch merges cleanly.
This patch adds no public classes.

dongjoon-hyun

The K8s IT failure looks relevant a little.

KubernetesSuite:
...
- Run in client mode. *** FAILED ***
  io.fabric8.kubernetes.client.KubernetesClientException: Failure executing: POST at: https://192.168.39.171:8443/api/v1/namespaces/2899456ab5484196859800707ad72210/pods. Message: No API token found for service account "default", retry after the token is automatically created and added to the service account. Received status: Status(apiVersion=v1, code=500, details=StatusDetails(causes=[], group=null, kind=serviceaccounts, name=create pod, retryAfterSeconds=1, uid=null, additionalProperties={}), kind=Status, message=No API token found for service account "default", retry after the token is automatically created and added to the service account, metadata=ListMeta(_continue=null, remainingItemCount=null, resourceVersion=null, selfLink=null, additionalProperties={}), reason=ServerTimeout, status=Failure, additionalProperties={}).
  at io.fabric8.kubernetes.client.dsl.base.OperationSupport.requestFailure(OperationSupport.java:589)
  at io.fabric8.kubernetes.client.dsl.base.OperationSupport.assertResponseCode(OperationSupport.java:528)
  at io.fabric8.kubernetes.client.dsl.base.OperationSupport.handleResponse(OperationSupport.java:492)
  at io.fabric8.kubernetes.client.dsl.base.OperationSupport.handleResponse(OperationSupport.java:451)
  at io.fabric8.kubernetes.client.dsl.base.OperationSupport.handleCreate(OperationSupport.java:252)
  at io.fabric8.kubernetes.client.dsl.base.BaseOperation.handleCreate(BaseOperation.java:844)
  at io.fabric8.kubernetes.client.dsl.base.BaseOperation.create(BaseOperation.java:352)
  at io.fabric8.kubernetes.client.dsl.base.BaseOperation.lambda$createNew$0(BaseOperation.java:365)
  at io.fabric8.kubernetes.api.model.DoneablePod.done(DoneablePod.java:26)
  at

laflechejonathan · 2020-09-30T09:40:16Z

I believe I've addressed your comments, thank you for the quick review!

@dongjoon-hyun, to answer your questions:

Done!
We are running against k8s 1.18.8
Correct, in spite of the declared compatibility, spark appears to run fine against k8s 1.18.x. That said, in our other java projects that rely on kubernetes-client, we routinely encounter subtle breaks and behaviour changes as we upgrade, so would like to keep spark as up to date as possible, so when break does occur, the version delta is smaller and less risky.

On the integration test failure, I'm fairly certain it's a flake. We used to run into this error when launching spark applications in production, and had to add functionality to block spark-submit until the service account token is created for a namespace. Would be good to add similar functionality to solve this flake, but should be out of scope of this PR.

SparkQA · 2020-09-30T10:09:08Z

Kubernetes integration test starting
URL: https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder-K8s/33886/

SparkQA · 2020-09-30T10:26:16Z

Kubernetes integration test status success
URL: https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder-K8s/33886/

SparkQA · 2020-09-30T12:01:48Z

Test build #129269 has finished for PR 29888 at commit 9aa0bdb.

This patch passes all tests.
This patch merges cleanly.
This patch adds no public classes.

dongjoon-hyun

+1, LGTM. Thank you, @laflechejonathan and @srowen .
Merged to master for Apache Spark 3.1.0 on December 2020.

dongjoon-hyun · 2020-10-01T02:03:07Z

Thank you again for your first contribution, @laflechejonathan .
I added you to the Apache Spark contributor group and assigned you to SPARK-33012.
Welcome to the Apache Spark community. :)

This PR aims to upgrade `kubernetes-client` library to track fabric8's declared compatibility for k8s 1.18.0: https://github.com/fabric8io/kubernetes-client#compatibility-matrix According to fabric8, 4.9.2 is incompatible with k8s 1.18.0. No. Not tested yet. Closes apache#29888 from laflechejonathan/jlf/fabric8Ugprade. Authored-by: jlafleche <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]>

* [SPARK-33012][BUILD][K8S] Upgrade fabric8 to 4.10.3 This PR aims to upgrade `kubernetes-client` library to track fabric8's declared compatibility for k8s 1.18.0: https://github.com/fabric8io/kubernetes-client#compatibility-matrix According to fabric8, 4.9.2 is incompatible with k8s 1.18.0. No. Not tested yet. Closes apache#29888 from laflechejonathan/jlf/fabric8Ugprade. Authored-by: jlafleche <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]> * Fix kubernetes tests Co-authored-by: jlafleche <[email protected]> Co-authored-by: Willi Raschkowski <[email protected]>

Upgrade fabric8 to 4.10.3

ee845e0

probot-autolabeler bot added BUILD KUBERNETES labels Sep 28, 2020

srowen reviewed Sep 28, 2020

View reviewed changes

dongjoon-hyun changed the title ~~[SPARK-33012] Upgrade fabric8 to 4.10.3~~ [SPARK-33012][BUILD][K8S] Upgrade fabric8 to 4.10.3 Sep 28, 2020

dongjoon-hyun reviewed Sep 28, 2020

View reviewed changes

resource-managers/kubernetes/core/pom.xml Show resolved Hide resolved

dongjoon-hyun reviewed Sep 28, 2020

View reviewed changes

dongjoon-hyun reviewed Sep 29, 2020

View reviewed changes

remove okhttp pin, regenerate manifests

9aa0bdb

dongjoon-hyun approved these changes Oct 1, 2020

View reviewed changes

dongjoon-hyun closed this in d75222d Oct 1, 2020

laflechejonathan mentioned this pull request Oct 1, 2020

[SPARK-33012][BUILD][K8S] Upgrade fabric8 to 4.10.3 palantir/spark#712

Merged

[SPARK-33012][BUILD][K8S] Upgrade fabric8 to 4.10.3 #29888

[SPARK-33012][BUILD][K8S] Upgrade fabric8 to 4.10.3 #29888

Uh oh!

Conversation

laflechejonathan commented Sep 28, 2020

What changes were proposed in this pull request?

Why are the changes needed?

Does this PR introduce any user-facing change?

How was this patch tested?

Uh oh!

srowen Sep 28, 2020

Choose a reason for hiding this comment

Uh oh!

dongjoon-hyun Sep 29, 2020

Choose a reason for hiding this comment

Uh oh!

laflechejonathan Sep 30, 2020 • 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!

srowen Sep 28, 2020

Choose a reason for hiding this comment

Uh oh!

laflechejonathan Sep 30, 2020

Choose a reason for hiding this comment

Uh oh!

dongjoon-hyun commented Sep 28, 2020

Uh oh!

Uh oh!

dongjoon-hyun left a comment • 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!

SparkQA commented Sep 28, 2020

Uh oh!

SparkQA commented Sep 28, 2020

Uh oh!

SparkQA commented Sep 28, 2020

Uh oh!

dongjoon-hyun left a comment • 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!

laflechejonathan commented Sep 30, 2020

Uh oh!

SparkQA commented Sep 30, 2020

Uh oh!

SparkQA commented Sep 30, 2020

Uh oh!

SparkQA commented Sep 30, 2020

Uh oh!

dongjoon-hyun left a comment

Choose a reason for hiding this comment

Uh oh!

dongjoon-hyun commented Oct 1, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

laflechejonathan Sep 30, 2020 •

edited

Loading

dongjoon-hyun left a comment •

edited

Loading

dongjoon-hyun left a comment •

edited

Loading

dongjoon-hyun commented Oct 1, 2020 •

edited

Loading