[
https://issues.apache.org/jira/browse/SPARK-60328?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
John Zhuge updated SPARK-60328:
-------------------------------
Description:
The cleanup in the HiveDDLSuite test "SPARK-34261: Avoid side effect if create
exists temporary function" removes the test jar only from the "default" bucket
of SparkContext.addedJars:
{code:scala}
.foreach(k => spark.sparkContext.addedJars.get("default").foreach(_.remove(k)))
{code}
But SQL commands run inside artifactManager.withResources, and
*spark.sql.artifact.isolation.enabled *defaults to true, so jars added through
SQL are stored under the session UUID. listJars() flattens all buckets, so a
jar left behind by an earlier test survives this cleanup and the next assertion
fails, depending on test order.
was:
The cleanup in the HiveDDLSuite test "SPARK-34261: Avoid side effect if create
exists temporary function" removes the test jar only from the "default" bucket
of SparkContext.addedJars:
{code:scala}
.foreach(k => spark.sparkContext.addedJars.get("default").foreach(_.remove(k)))
{code}
But SQL commands run inside artifactManager.withResources, and
*spark.sql.artifact.isolation.enabled *defaults to true, so jars added through
SQL are stored under the session UUID. listJars() flattens all buckets, so a
jar left behind by an earlier test survives this cleanup and the next assertion
fails, depending on test order.
Fix: remove the jar from every bucket. Test-only change.
{code:scala}
.foreach(k => spark.sparkContext.addedJars.values.foreach(_.remove(k)))
{code}
> HiveDDLSuite SPARK-34261 test fails depending on test order
> -----------------------------------------------------------
>
> Key: SPARK-60328
> URL: https://issues.apache.org/jira/browse/SPARK-60328
> Project: Spark
> Issue Type: Test
> Components: SQL
> Affects Versions: 4.2.0
> Reporter: John Zhuge
> Priority: Minor
> Labels: pull-request-available
>
> The cleanup in the HiveDDLSuite test "SPARK-34261: Avoid side effect if
> create exists temporary function" removes the test jar only from the
> "default" bucket of SparkContext.addedJars:
> {code:scala}
> .foreach(k =>
> spark.sparkContext.addedJars.get("default").foreach(_.remove(k)))
> {code}
> But SQL commands run inside artifactManager.withResources, and
> *spark.sql.artifact.isolation.enabled *defaults to true, so jars added
> through SQL are stored under the session UUID. listJars() flattens all
> buckets, so a jar left behind by an earlier test survives this cleanup and
> the next assertion fails, depending on test order.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
