Skip to content

Conversation

@andrewor14
Copy link
Contributor

The REST server is not actually used in most tests and so we can disable it. It is a source of flakiness because it tries to bind to a specific port in vain. There was also some code that avoided the shuffle service in tests. This is actually not necessary because the shuffle service is already off by default.

@JoshRosen
Copy link
Contributor

Are there any tests where we need to explicitly re-enable this configuration or is this feature completely untested?

@andrewor14
Copy link
Contributor Author

It is tested in StandaloneRestSubmissionSuite but is mocked out there and doesn't go through the standalone Master.

@SparkQA
Copy link

SparkQA commented Aug 11, 2015

Test build #40343 timed out for PR 8084 at commit f86f2be after a configured wait of 175m.

@JoshRosen
Copy link
Contributor

Jenkins, retest this please.

@SparkQA
Copy link

SparkQA commented Aug 11, 2015

Test build #40411 timed out for PR 8084 at commit f86f2be after a configured wait of 175m.

@andrewor14
Copy link
Contributor Author

retest this please

@andrewor14
Copy link
Contributor Author

retest this please?

@SparkQA
Copy link

SparkQA commented Aug 11, 2015

Test build #40475 timed out for PR 8084 at commit f86f2be after a configured wait of 175m.

There was some code that also avoided the shuffle service in tests.
This is actually not necessary because the shuffle service is
already off by default.
@andrewor14 andrewor14 force-pushed the fix-master-suite-again branch from f86f2be to 026f628 Compare August 11, 2015 21:27
@andrewor14
Copy link
Contributor Author

retest this please

@SparkQA
Copy link

SparkQA commented Aug 12, 2015

Test build #1456 has finished for PR 8084 at commit 026f628.

  • This patch fails Spark unit tests.
  • This patch does not merge cleanly.
  • This patch adds no public classes.

@andrewor14
Copy link
Contributor Author

retest this please

@SparkQA
Copy link

SparkQA commented Aug 12, 2015

Test build #40523 timed out for PR 8084 at commit 026f628 after a configured wait of 175m.

@JoshRosen
Copy link
Contributor

Jenkins, retest this please.

@SparkQA
Copy link

SparkQA commented Aug 12, 2015

Test build #1465 has finished for PR 8084 at commit 026f628.

  • This patch fails Spark unit tests.
  • This patch merges cleanly.
  • This patch adds no public classes.

@JoshRosen
Copy link
Contributor

Jenkins, retest this please.

@JoshRosen
Copy link
Contributor

Wow, this has been the unluckiest PR when it comes to flakiness :(

@SparkQA
Copy link

SparkQA commented Aug 12, 2015

Test build #1466 has finished for PR 8084 at commit 026f628.

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

@JoshRosen
Copy link
Contributor

LGTM, so I'm going to merge this to master and branch-1.5.

asfgit pushed a commit that referenced this pull request Aug 12, 2015
The REST server is not actually used in most tests and so we can disable it. It is a source of flakiness because it tries to bind to a specific port in vain. There was also some code that avoided the shuffle service in tests. This is actually not necessary because the shuffle service is already off by default.

Author: Andrew Or <andrew@databricks.com>

Closes #8084 from andrewor14/fix-master-suite-again.

(cherry picked from commit ca8f70e)
Signed-off-by: Josh Rosen <joshrosen@databricks.com>
@asfgit asfgit closed this in ca8f70e Aug 12, 2015
@andrewor14 andrewor14 deleted the fix-master-suite-again branch August 12, 2015 04:36
@SparkQA
Copy link

SparkQA commented Aug 12, 2015

Test build #40575 has finished for PR 8084 at commit 026f628.

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

@SparkQA
Copy link

SparkQA commented Aug 12, 2015

Test build #40580 has finished for PR 8084 at commit 026f628.

  • This patch fails PySpark unit tests.
  • This patch merges cleanly.
  • This patch adds no public classes.

@JoshRosen
Copy link
Contributor

@andrewor14 do you think that this is also a candidate for backport to 1.4.x? I think that I've seen similar MasterSuite flakiness there as well.

@andrewor14
Copy link
Contributor Author

Yeah, I think it doesn't hurt.

CodingCat pushed a commit to CodingCat/spark that referenced this pull request Aug 17, 2015
The REST server is not actually used in most tests and so we can disable it. It is a source of flakiness because it tries to bind to a specific port in vain. There was also some code that avoided the shuffle service in tests. This is actually not necessary because the shuffle service is already off by default.

Author: Andrew Or <andrew@databricks.com>

Closes apache#8084 from andrewor14/fix-master-suite-again.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants