You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@spark.apache.org by "Hyukjin Kwon (Jira)" <ji...@apache.org> on 2021/07/20 02:42:00 UTC

[jira] [Comment Edited] (SPARK-36218) Flaky Test: TPC-DS in PR builder

    [ https://issues.apache.org/jira/browse/SPARK-36218?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17383707#comment-17383707 ] 

Hyukjin Kwon edited comment on SPARK-36218 at 7/20/21, 2:41 AM:
----------------------------------------------------------------

cc [~maropu], [~cloud_fan], [~dongjoon] FYI.

Actually, I faced this issue in our internal repo a while ago, and just added a hacky fix by adding an explicit GC:

{code}
  if (tpcdsDataPath.nonEmpty) {
    tpcdsQueries
      .foreach { name =>
      val queryString = resourceToString(s"tpcds/$name.sql",
        classLoader = Thread.currentThread().getContextClassLoader)
      test(name) {
        + // SPARK-36218: workaround to prevent unexpected failure related to resource usage.
        + System.gc()
        val goldenFile = new File(s"$baseResourcePath/v1_4", s"$name.sql.out")
        runQuery(queryString, goldenFile)
      }
    }
    tpcdsQueriesV2_7_0
      .foreach { name =>
      val queryString = resourceToString(s"tpcds-v2.7.0/$name.sql",
        classLoader = Thread.currentThread().getContextClassLoader)
      test(s"$name-v2.7") {
        + // SPARK-36218: workaround to prevent unexpected failure related to resource usage.
        + System.gc()
        val goldenFile = new File(s"$baseResourcePath/v2_7", s"$name.sql.out")
        runQuery(queryString, goldenFile)
      }
    }
  } else {
    ignore("skipped because env `SPARK_TPCDS_DATA` is not set") {}
  }
}
{code}


was (Author: hyukjin.kwon):
cc [~maropu], [~cloud_fan], [~dongjoon] FYI.

Actually, I faced this issue in our internal repo a while ago, and just added a hacky fix by adding an explicit GC:

{code}
  if (tpcdsDataPath.nonEmpty) {
    tpcdsQueries
      .filter(_ != "q95") // TODO(SC-75125)
      .filter(_ != "q75") // TODO(SC-75127)
      .filter(_ != "q64") // TODO(SC-75126)
      .foreach { name =>
      val queryString = resourceToString(s"tpcds/$name.sql",
        classLoader = Thread.currentThread().getContextClassLoader)
      test(name) {
        + // SPARK-36218: workaround to prevent unexpected failure related to resource usage.
        + System.gc()
        val goldenFile = new File(s"$baseResourcePath/v1_4", s"$name.sql.out")
        runQuery(queryString, goldenFile)
      }
    }
    tpcdsQueriesV2_7_0
      .filter(_ != "q95") // TODO(SC-75125)
      .filter(_ != "q75") // TODO(SC-75127)
      .filter(_ != "q64") // TODO(SC-75126)
      .foreach { name =>
      val queryString = resourceToString(s"tpcds-v2.7.0/$name.sql",
        classLoader = Thread.currentThread().getContextClassLoader)
      test(s"$name-v2.7") {
        + // SPARK-36218: workaround to prevent unexpected failure related to resource usage.
        + System.gc()
        val goldenFile = new File(s"$baseResourcePath/v2_7", s"$name.sql.out")
        runQuery(queryString, goldenFile)
      }
    }
  } else {
    ignore("skipped because env `SPARK_TPCDS_DATA` is not set") {}
  }
}
{code}

> Flaky Test: TPC-DS in PR builder
> --------------------------------
>
>                 Key: SPARK-36218
>                 URL: https://issues.apache.org/jira/browse/SPARK-36218
>             Project: Spark
>          Issue Type: Test
>          Components: SQL, Tests
>    Affects Versions: 3.0.3, 3.1.2, 3.2.0, 3.3.0
>            Reporter: Hyukjin Kwon
>            Priority: Major
>
> {code}
> [info] - q1 (9 seconds, 603 milliseconds)
> [info] - q2 (5 seconds, 860 milliseconds)
> [info] - q3 (1 second, 777 milliseconds)
> [info] - q4 (31 seconds, 951 milliseconds)
> [info] - q5 (4 seconds, 561 milliseconds)
> [info] - q7 (2 seconds, 471 milliseconds)
> [info] - q8 (2 seconds, 74 milliseconds)
> [info] - q9 (4 seconds, 402 milliseconds)
> [info] - q10 (4 seconds, 618 milliseconds)
> /home/runner/work/spark/spark/build/sbt-launch-lib.bash: line 77:  1659 Killed                  "$@"
> Error: Process completed with exit code 137.
> {code}
> It dies in the middle: https://github.com/apache/spark/runs/3109502701



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org