[jira] [Comment Edited] (SPARK-9487) Use the same num. worker threads in Scala/Python unit tests
[ https://issues.apache.org/jira/browse/SPARK-9487?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15680080#comment-15680080 ] Saikat Kanjilal edited comment on SPARK-9487 at 11/19/16 11:59 PM: --- Ok guess I spoke too soon :), onto the next set of challenges, jenkins build report is here: https://amplab.cs.berkeley.edu/jenkins//job/SparkPullRequestBuilder/68897/ I ran each of these tests individually as well as together as a suite locally and they all passed, any ideas on how to address these? was (Author: kanjilal): Ok guess I spoke too soon :), onto the next set of challenges, jenkins build report is here: https://amplab.cs.berkeley.edu/jenkins//job/SparkPullRequestBuilder/68897/ I ran each of these tests individually as well as together as a suite and they all passed, any ideas on how to address these? > Use the same num. worker threads in Scala/Python unit tests > --- > > Key: SPARK-9487 > URL: https://issues.apache.org/jira/browse/SPARK-9487 > Project: Spark > Issue Type: Improvement > Components: PySpark, Spark Core, SQL, Tests >Affects Versions: 1.5.0 >Reporter: Xiangrui Meng > Labels: starter > Attachments: ContextCleanerSuiteResults, HeartbeatReceiverSuiteResults > > > In Python we use `local[4]` for unit tests, while in Scala/Java we use > `local[2]` and `local` for some unit tests in SQL, MLLib, and other > components. If the operation depends on partition IDs, e.g., random number > generator, this will lead to different result in Python and Scala/Java. It > would be nice to use the same number in all unit tests. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Comment Edited] (SPARK-9487) Use the same num. worker threads in Scala/Python unit tests
[ https://issues.apache.org/jira/browse/SPARK-9487?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15658479#comment-15658479 ] Saikat Kanjilal edited comment on SPARK-9487 at 11/11/16 11:20 PM: --- ok so I've spent the last hour or so doing deeper investigations into the failures, I used this https://amplab.cs.berkeley.edu/jenkins//job/SparkPullRequestBuilder/68531/ as a point of reference, listed below is what I found LogisticRegressionSuite (passed successfully on both my master and feature branches) OneVsRestSuite(passed successfully on both my master and feature branches) DataFrameStatSuite (passed successfully on both my master and feature branches) DataFrameSuite (passed successfully on both my master and feature branches) SQLQueryTestSuite (passed successfully on both my master and feature branches) ForeachSinkSuite (passed successfully on both my master and feature branches) JavaAPISuite (failed on both my master and feature branches) The master branch does not have any code changes from me and the feature branch of course does I am running individual tests by issuing commands like the following from the root directory based on the documentation: ./build/mvn test -P... -DwildcardSuites=none -Dtest=org.apache.spark.streaming.JavaAPISuite Therefore my conclusion so far based on the above jenkins report is that my changes have not introduced any new failures that were not already there, [~srowen] please let me know if my methodology is off anywhere was (Author: kanjilal): ok so I've spent the last hour or so doing deeper investigations into the failures, I used this https://amplab.cs.berkeley.edu/jenkins//job/SparkPullRequestBuilder/68531/ as a point of reference, listed below is what I found java/scala test my master branch my feature branch LogisticRegressionSuite success success OneVsRestSuite success success DataFrameStatSuite success success DataFrameSuite success success SQLQueryTestSuitesuccess success ForeachSinkSuite success success JavaAPISuite failure failure The master branch does not have any code changes from me and the feature branch of course does I am running individual tests by issuing commands like the following from the root directory based on the documentation: ./build/mvn test -P... -DwildcardSuites=none -Dtest=org.apache.spark.streaming.JavaAPISuite Therefore my conclusion so far based on the above jenkins report is that my changes have not introduced any new failures that were not already there, [~srowen] please let me know if my methodology is off anywhere > Use the same num. worker threads in Scala/Python unit tests > --- > > Key: SPARK-9487 > URL: https://issues.apache.org/jira/browse/SPARK-9487 > Project: Spark > Issue Type: Improvement > Components: PySpark, Spark Core, SQL, Tests >Affects Versions: 1.5.0 >Reporter: Xiangrui Meng > Labels: starter > Attachments: ContextCleanerSuiteResults, HeartbeatReceiverSuiteResults > > > In Python we use `local[4]` for unit tests, while in Scala/Java we use > `local[2]` and `local` for some unit tests in SQL, MLLib, and other > components. If the operation depends on partition IDs, e.g., random number > genera
[jira] [Comment Edited] (SPARK-9487) Use the same num. worker threads in Scala/Python unit tests
[ https://issues.apache.org/jira/browse/SPARK-9487?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15651806#comment-15651806 ] Saikat Kanjilal edited comment on SPARK-9487 at 11/9/16 7:22 PM: - Ok , for some odd reason my local branch had the changes but weren't committed, PR is here: https://github.com/skanjila/spark/commit/ec0b2a81dc8362e84e70457873560d997a7cb244 I added the change to local[4] to both streaming as well as repl, based on what I'm seeing locally all Java/Scala changes should be accounted for and unit tests pass, the only exception is the code inside spark examples in PageViewStream.scala, should I change this, seems like it doesn't belong as part of the unit tests. My next TODOs: 1) Change the example code if it makes sense in PageViewStream 2) Start the code changes to fix the python unit tests Let me know thoughts or concerns. was (Author: kanjilal): Ok , for some odd reason my local branch had the changes but weren't committed, PR is here: https://github.com/skanjila/spark/commit/ec0b2a81dc8362e84e70457873560d997a7cb244 I added the change to local[4] to both streaming as well as repl, based on what I'm seeing locally all Java/Scala changes should be accounted for except for spark examples with the code inside PageViewStream.scala, should I change this, seems like it doesn't belong as part of the unit tests. My next TODOs: 1) Change the example code if it makes sense in PageViewStream 2) Start the code changes to fix the python unit tests Let me know thoughts or concerns. > Use the same num. worker threads in Scala/Python unit tests > --- > > Key: SPARK-9487 > URL: https://issues.apache.org/jira/browse/SPARK-9487 > Project: Spark > Issue Type: Improvement > Components: PySpark, Spark Core, SQL, Tests >Affects Versions: 1.5.0 >Reporter: Xiangrui Meng > Labels: starter > Attachments: ContextCleanerSuiteResults, HeartbeatReceiverSuiteResults > > > In Python we use `local[4]` for unit tests, while in Scala/Java we use > `local[2]` and `local` for some unit tests in SQL, MLLib, and other > components. If the operation depends on partition IDs, e.g., random number > generator, this will lead to different result in Python and Scala/Java. It > would be nice to use the same number in all unit tests. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Comment Edited] (SPARK-9487) Use the same num. worker threads in Scala/Python unit tests
[ https://issues.apache.org/jira/browse/SPARK-9487?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15620613#comment-15620613 ] Saikat Kanjilal edited comment on SPARK-9487 at 10/30/16 9:46 PM: -- [~srowen] Yes I read through that link and adjusted the PR title, however please do let me know if I can proceed adding more to this PR including python and other parts of the codebase. was (Author: kanjilal): [~srowen] Yes I read through that link and adjusted the PR title, I will Jenkins test this next, however please do let me know if I can proceed adding more to this PR including python and other parts of the codebase. > Use the same num. worker threads in Scala/Python unit tests > --- > > Key: SPARK-9487 > URL: https://issues.apache.org/jira/browse/SPARK-9487 > Project: Spark > Issue Type: Improvement > Components: PySpark, Spark Core, SQL, Tests >Affects Versions: 1.5.0 >Reporter: Xiangrui Meng > Labels: starter > Attachments: ContextCleanerSuiteResults, HeartbeatReceiverSuiteResults > > > In Python we use `local[4]` for unit tests, while in Scala/Java we use > `local[2]` and `local` for some unit tests in SQL, MLLib, and other > components. If the operation depends on partition IDs, e.g., random number > generator, this will lead to different result in Python and Scala/Java. It > would be nice to use the same number in all unit tests. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Comment Edited] (SPARK-9487) Use the same num. worker threads in Scala/Python unit tests
[ https://issues.apache.org/jira/browse/SPARK-9487?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15620613#comment-15620613 ] Saikat Kanjilal edited comment on SPARK-9487 at 10/30/16 9:40 PM: -- [~srowen] Yes I read through that link and adjusted the PR title, I will Jenkins test this next, however please do let me know if I can proceed adding more to this PR including python and other parts of the codebase. was (Author: kanjilal): [~srowen] Yes I read through that and adjusted the PR title, I will Jenkins test this next, however please do let me know if I can proceed adding more to this PR including python and other parts of the codebase. > Use the same num. worker threads in Scala/Python unit tests > --- > > Key: SPARK-9487 > URL: https://issues.apache.org/jira/browse/SPARK-9487 > Project: Spark > Issue Type: Improvement > Components: PySpark, Spark Core, SQL, Tests >Affects Versions: 1.5.0 >Reporter: Xiangrui Meng > Labels: starter > Attachments: ContextCleanerSuiteResults, HeartbeatReceiverSuiteResults > > > In Python we use `local[4]` for unit tests, while in Scala/Java we use > `local[2]` and `local` for some unit tests in SQL, MLLib, and other > components. If the operation depends on partition IDs, e.g., random number > generator, this will lead to different result in Python and Scala/Java. It > would be nice to use the same number in all unit tests. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Comment Edited] (SPARK-9487) Use the same num. worker threads in Scala/Python unit tests
[ https://issues.apache.org/jira/browse/SPARK-9487?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15587627#comment-15587627 ] Saikat Kanjilal edited comment on SPARK-9487 at 10/19/16 4:24 AM: -- [~holdenk] finally getting time to look at this, so I am starting small, I made the change inside ContextCleanerSuite and HeartbeatReceiverSuite from local[2] tp local[4], per the documentation here (http://spark.apache.org/docs/latest/building-spark.html#specifying-the-hadoop-version) I ran mvn -P hadoop2 -Dsuites= org.apache.spark.HeartbeatReceiverSuite test--- looks like everything worked I then ran mvn -P hadoop2 -Dsuites= org.apache.spark.ContextCleanerSuite test-- looks like everything worked as well See the attachments and let me know if this is the right process to run single unit tests, I'll start making changes to the other Suites , how would you like to see the output, should I just have attachments or just do a pull request from the new branch that I created? Thanks PS Another question, running single unit tests like this takes forever, are there flags I can set to speed up the builds, even on my 15 inch macbook pro with SSD the builds shouldnt take this long :(. Let me know next steps to get this into a PR. was (Author: kanjilal): [~holdenk] finally getting time to look at this, so I am starting small, I made the change inside ContextCleanerSuite and HeartbeatReceiverSuite from local[2] tp local[4], per the documentation here (http://spark.apache.org/docs/latest/building-spark.html#specifying-the-hadoop-version) I ran mvn -P hadoop2 -Dsuites= org.apache.spark.HeartbeatReceiverSuite test--- looks like everything worked I then ran mvn -P hadoop2 -Dsuites= org.apache.spark.ContextCleanerSuite test-- looks like everything worked as well See the attachments and let me know if this is this is the right process to run single unit tests, if not I'll start making changes to the other Suites , how would you like to see the output, should I just have attachments or just do a pull request from the new branch that I created? Thanks PS Another question, running single unit tests like this takes forever, are there flags I can set to speed up the builds, even on my 15 inch macbook pro with SSD the builds shouldnt take this long :(. Let me know next steps to get this into a PR. > Use the same num. worker threads in Scala/Python unit tests > --- > > Key: SPARK-9487 > URL: https://issues.apache.org/jira/browse/SPARK-9487 > Project: Spark > Issue Type: Improvement > Components: PySpark, Spark Core, SQL, Tests >Affects Versions: 1.5.0 >Reporter: Xiangrui Meng > Labels: starter > Attachments: ContextCleanerSuiteResults, HeartbeatReceiverSuiteResults > > > In Python we use `local[4]` for unit tests, while in Scala/Java we use > `local[2]` and `local` for some unit tests in SQL, MLLib, and other > components. If the operation depends on partition IDs, e.g., random number > generator, this will lead to different result in Python and Scala/Java. It > would be nice to use the same number in all unit tests. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Comment Edited] (SPARK-9487) Use the same num. worker threads in Scala/Python unit tests
[ https://issues.apache.org/jira/browse/SPARK-9487?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15587627#comment-15587627 ] Saikat Kanjilal edited comment on SPARK-9487 at 10/19/16 4:24 AM: -- [~holdenk] finally getting time to look at this, so I am starting small, I made the change inside ContextCleanerSuite and HeartbeatReceiverSuite from local[2] tp local[4], per the documentation here (http://spark.apache.org/docs/latest/building-spark.html#specifying-the-hadoop-version) I ran mvn -P hadoop2 -Dsuites= org.apache.spark.HeartbeatReceiverSuite test--- looks like everything worked I then ran mvn -P hadoop2 -Dsuites= org.apache.spark.ContextCleanerSuite test-- looks like everything worked as well See the attachments and let me know if this is this is the right process to run single unit tests, if not I'll start making changes to the other Suites , how would you like to see the output, should I just have attachments or just do a pull request from the new branch that I created? Thanks PS Another question, running single unit tests like this takes forever, are there flags I can set to speed up the builds, even on my 15 inch macbook pro with SSD the builds shouldnt take this long :(. Let me know next steps to get this into a PR. was (Author: kanjilal): [~holdenk] finally getting time to look at this, so I am starting small, I made the change inside ContextCleanerSuite and HeartbeatReceiverSuite from local[2] tp local[4], per the documentation here (http://spark.apache.org/docs/latest/building-spark.html#specifying-the-hadoop-version) I ran mvn -Phadoop2 -Dsuites=org.apache.spark.HeartbeatReceiverSuite test--- looks like everything worked I then ran mvn -Phadoop2 -Dsuites=org.apache.spark.ContextCleanerSuite test-- looks like everything worked as well See the attachments and let me know if this is not the right process to run single unit tests, if not I'll start making changes to the other Suites , how would you like to see the output, should I just have attachments or just do a pull request from the new branch that I created? Thanks PS Another question, running single unit tests like this takes forever, are there flags I can set to speed up the builds, even on my 15 inch macbook pro with SSD the builds shouldnt take this long :(. Let me know next steps to get this into a PR. > Use the same num. worker threads in Scala/Python unit tests > --- > > Key: SPARK-9487 > URL: https://issues.apache.org/jira/browse/SPARK-9487 > Project: Spark > Issue Type: Improvement > Components: PySpark, Spark Core, SQL, Tests >Affects Versions: 1.5.0 >Reporter: Xiangrui Meng > Labels: starter > Attachments: ContextCleanerSuiteResults, HeartbeatReceiverSuiteResults > > > In Python we use `local[4]` for unit tests, while in Scala/Java we use > `local[2]` and `local` for some unit tests in SQL, MLLib, and other > components. If the operation depends on partition IDs, e.g., random number > generator, this will lead to different result in Python and Scala/Java. It > would be nice to use the same number in all unit tests. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Comment Edited] (SPARK-9487) Use the same num. worker threads in Scala/Python unit tests
[ https://issues.apache.org/jira/browse/SPARK-9487?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14941961#comment-14941961 ] Evan Chen edited comment on SPARK-9487 at 10/6/15 11:24 PM: Hey [~mengxr], What would be the preferred num. worker threads? Should we set all of them to local[2] to stay consistent with the Scala/Java side? Thanks was (Author: evanchen92): Hey Xiangrui, What would be the preferred num. worker threads? Should we set all of them to local[2] to stay consistent with the Scala/Java side? Thanks > Use the same num. worker threads in Scala/Python unit tests > --- > > Key: SPARK-9487 > URL: https://issues.apache.org/jira/browse/SPARK-9487 > Project: Spark > Issue Type: Improvement > Components: PySpark, Spark Core, SQL, Tests >Affects Versions: 1.5.0 >Reporter: Xiangrui Meng > > In Python we use `local[4]` for unit tests, while in Scala/Java we use > `local[2]` and `local` for some unit tests in SQL, MLLib, and other > components. If the operation depends on partition IDs, e.g., random number > generator, this will lead to different result in Python and Scala/Java. It > would be nice to use the same number in all unit tests. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org