You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@spark.apache.org by "Takeshi Yamamuro (Jira)" <ji...@apache.org> on 2020/08/10 11:49:00 UTC
[jira] [Comment Edited] (SPARK-29767) Core dump happening on
executors while doing simple union of Data Frames
[ https://issues.apache.org/jira/browse/SPARK-29767?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17174263#comment-17174263 ]
Takeshi Yamamuro edited comment on SPARK-29767 at 8/10/20, 11:48 AM:
---------------------------------------------------------------------
I checked the given query and I think your union query looks incorrect. All the branches (master/branch-3.0/branch-2.4) have the same error;
{code:java}
: org.apache.spark.sql.AnalysisException: Union can only be performed on tables with the same number of columns, but the first table has 5 columns and the second table has 20 columns;; {code}
Actually, the schemas are different;
{code:java}
>>> base_df.printSchema()
root
|-- hour: long (nullable = true)
|-- title: string (nullable = true)
|-- __deleted: string (nullable = true)
|-- status: string (nullable = true)
|-- transformationid: string (nullable = true)
|-- roomid: string (nullable = true)
|-- day: long (nullable = true)
|-- notes: string (nullable = true)
|-- nunitsfromaudit: long (nullable = true)
|-- ts_ms: long (nullable = true)
|-- liability: string (nullable = true)
|-- _class: string (nullable = true)
|-- month: long (nullable = true)
|-- updatedate: struct (nullable = true)
| |-- date: long (nullable = true)
|-- _id: struct (nullable = true)
| |-- oid: string (nullable = true)
|-- year: long (nullable = true)
|-- item: struct (nullable = true)
| |-- name: string (nullable = true)
| |-- brandname: string (nullable = true)
| |-- perunitpricefromaudit: struct (nullable = true)
| | |-- currency: string (nullable = true)
| | |-- amount: string (nullable = true)
| |-- actualperunitprice: struct (nullable = true)
| | |-- currency: string (nullable = true)
| | |-- amount: string (nullable = true)
| |-- category: string (nullable = true)
| |-- itemtype: string (nullable = true)
| |-- roomamenityid: long (nullable = true)
|-- createddate: struct (nullable = true)
| |-- date: long (nullable = true)
|-- actualunits: long (nullable = true)
|-- description: string (nullable = true)
>>> inc_df.printSchema()
root
|-- _id: struct (nullable = true)
| |-- oid: string (nullable = true)
|-- _class: string (nullable = true)
|-- roomid: string (nullable = true)
|-- item: struct (nullable = true)
| |-- name: string (nullable = true)
| |-- brandname: string (nullable = true)
| |-- perunitpricefromaudit: struct (nullable = true)
| | |-- currency: string (nullable = true)
| | |-- amount: string (nullable = true)
| |-- actualperunitprice: struct (nullable = true)
| | |-- currency: string (nullable = true)
| | |-- amount: string (nullable = true)
| |-- category: string (nullable = true)
| |-- itemtype: string (nullable = true)
| |-- roomamenityid: long (nullable = true)
|-- inserted_at: string (nullable = true)
{code}
NOTE: It is the best to show us a simpler query to reproduce a issue. So, if possible, its better to remove unnecessary parts, e.g., columns, for saving developer's time.
was (Author: maropu):
I checked the given query and I think your union query looks incorrect. All the branches (master/branch-3.0/branch-2.4) have the same error;
{code:java}
: org.apache.spark.sql.AnalysisException: Union can only be performed on tables with the same number of columns, but the first table has 5 columns and the second table has 20 columns;; {code}
Actually, the schema are different;
{code:java}
>>> base_df.printSchema()
root
|-- hour: long (nullable = true)
|-- title: string (nullable = true)
|-- __deleted: string (nullable = true)
|-- status: string (nullable = true)
|-- transformationid: string (nullable = true)
|-- roomid: string (nullable = true)
|-- day: long (nullable = true)
|-- notes: string (nullable = true)
|-- nunitsfromaudit: long (nullable = true)
|-- ts_ms: long (nullable = true)
|-- liability: string (nullable = true)
|-- _class: string (nullable = true)
|-- month: long (nullable = true)
|-- updatedate: struct (nullable = true)
| |-- date: long (nullable = true)
|-- _id: struct (nullable = true)
| |-- oid: string (nullable = true)
|-- year: long (nullable = true)
|-- item: struct (nullable = true)
| |-- name: string (nullable = true)
| |-- brandname: string (nullable = true)
| |-- perunitpricefromaudit: struct (nullable = true)
| | |-- currency: string (nullable = true)
| | |-- amount: string (nullable = true)
| |-- actualperunitprice: struct (nullable = true)
| | |-- currency: string (nullable = true)
| | |-- amount: string (nullable = true)
| |-- category: string (nullable = true)
| |-- itemtype: string (nullable = true)
| |-- roomamenityid: long (nullable = true)
|-- createddate: struct (nullable = true)
| |-- date: long (nullable = true)
|-- actualunits: long (nullable = true)
|-- description: string (nullable = true)
>>> inc_df.printSchema()
root
|-- _id: struct (nullable = true)
| |-- oid: string (nullable = true)
|-- _class: string (nullable = true)
|-- roomid: string (nullable = true)
|-- item: struct (nullable = true)
| |-- name: string (nullable = true)
| |-- brandname: string (nullable = true)
| |-- perunitpricefromaudit: struct (nullable = true)
| | |-- currency: string (nullable = true)
| | |-- amount: string (nullable = true)
| |-- actualperunitprice: struct (nullable = true)
| | |-- currency: string (nullable = true)
| | |-- amount: string (nullable = true)
| |-- category: string (nullable = true)
| |-- itemtype: string (nullable = true)
| |-- roomamenityid: long (nullable = true)
|-- inserted_at: string (nullable = true)
{code}
NOTE: It is the best to show us a simpler query to reproduce a issue. So, if possible, its better to remove unnecessary parts, e.g., columns, for saving developer's time.
> Core dump happening on executors while doing simple union of Data Frames
> ------------------------------------------------------------------------
>
> Key: SPARK-29767
> URL: https://issues.apache.org/jira/browse/SPARK-29767
> Project: Spark
> Issue Type: Bug
> Components: PySpark, Spark Core
> Affects Versions: 2.4.4
> Environment: AWS EMR 5.27.0, Spark 2.4.4
> Reporter: Udit Mehrotra
> Priority: Major
> Attachments: coredump.zip, hs_err_pid13885.log, part-00000-0189b5c2-7f7b-4d0e-bdb8-506380253597-c000.snappy.parquet, test.py
>
>
> Running a union operation on two DataFrames through both Scala Spark Shell and PySpark, resulting in executor contains doing a *core dump* and existing with Exit code 134.
> The trace from the *Driver*:
> {noformat}
> Container exited with a non-zero exit code 134
> .
> 19/11/06 02:21:35 ERROR TaskSetManager: Task 0 in stage 2.0 failed 4 times; aborting job
> org.apache.spark.SparkException: Job aborted due to stage failure: Task 0 in stage 2.0 failed 4 times, most recent failure: Lost task 0.3 in stage 2.0 (TID 5, ip-172-30-6-79.ec2.internal, executor 11): ExecutorLostFailure (executor 11 exited caused by one of the running tasks) Reason: Container from a bad node: container_1572981097605_0021_01_000077 on host: ip-172-30-6-79.ec2.internal. Exit status: 134. Diagnostics: Exception from container-launch.
> Container id: container_1572981097605_0021_01_000077
> Exit code: 134
> Exception message: /bin/bash: line 1: 12611 Aborted LD_LIBRARY_PATH="/usr/lib/hadoop/lib/native:/usr/lib/hadoop-lzo/lib/native:::/usr/lib/hadoop-lzo/lib/native:/usr/lib/hadoop/lib/native::/usr/lib/hadoop-lzo/lib/native:/usr/lib/hadoop/lib/native:/usr/lib/hadoop-lzo/lib/native:/usr/lib/hadoop/lib/native" /usr/lib/jvm/java-openjdk/bin/java -server -Xmx2743m '-verbose:gc' '-XX:+PrintGCDetails' '-XX:+PrintGCDateStamps' '-XX:+UseConcMarkSweepGC' '-XX:CMSInitiatingOccupancyFraction=70' '-XX:MaxHeapFreeRatio=70' '-XX:+CMSClassUnloadingEnabled' '-XX:OnOutOfMemoryError=kill -9 %p' -Djava.io.tmpdir=/mnt1/yarn/usercache/hadoop/appcache/application_1572981097605_0021/container_1572981097605_0021_01_000077/tmp '-Dspark.history.ui.port=18080' '-Dspark.driver.port=42267' -Dspark.yarn.app.container.log.dir=/var/log/hadoop-yarn/containers/application_1572981097605_0021/container_1572981097605_0021_01_000077 org.apache.spark.executor.CoarseGrainedExecutorBackend --driver-url spark://CoarseGrainedScheduler@ip-172-30-6-103.ec2.internal:42267 --executor-id 11 --hostname ip-172-30-6-79.ec2.internal --cores 2 --app-id application_1572981097605_0021 --user-class-path file:/mnt1/yarn/usercache/hadoop/appcache/application_1572981097605_0021/container_1572981097605_0021_01_000077/__app__.jar > /var/log/hadoop-yarn/containers/application_1572981097605_0021/container_1572981097605_0021_01_000077/stdout 2> /var/log/hadoop-yarn/containers/application_1572981097605_0021/container_1572981097605_0021_01_000077/stderrStack trace: ExitCodeException exitCode=134: /bin/bash: line 1: 12611 Aborted LD_LIBRARY_PATH="/usr/lib/hadoop/lib/native:/usr/lib/hadoop-lzo/lib/native:::/usr/lib/hadoop-lzo/lib/native:/usr/lib/hadoop/lib/native::/usr/lib/hadoop-lzo/lib/native:/usr/lib/hadoop/lib/native:/usr/lib/hadoop-lzo/lib/native:/usr/lib/hadoop/lib/native" /usr/lib/jvm/java-openjdk/bin/java -server -Xmx2743m '-verbose:gc' '-XX:+PrintGCDetails' '-XX:+PrintGCDateStamps' '-XX:+UseConcMarkSweepGC' '-XX:CMSInitiatingOccupancyFraction=70' '-XX:MaxHeapFreeRatio=70' '-XX:+CMSClassUnloadingEnabled' '-XX:OnOutOfMemoryError=kill -9 %p' -Djava.io.tmpdir=/mnt1/yarn/usercache/hadoop/appcache/application_1572981097605_0021/container_1572981097605_0021_01_000077/tmp '-Dspark.history.ui.port=18080' '-Dspark.driver.port=42267' -Dspark.yarn.app.container.log.dir=/var/log/hadoop-yarn/containers/application_1572981097605_0021/container_1572981097605_0021_01_000077 org.apache.spark.executor.CoarseGrainedExecutorBackend --driver-url spark://CoarseGrainedScheduler@ip-172-30-6-103.ec2.internal:42267 --executor-id 11 --hostname ip-172-30-6-79.ec2.internal --cores 2 --app-id application_1572981097605_0021 --user-class-path file:/mnt1/yarn/usercache/hadoop/appcache/application_1572981097605_0021/container_1572981097605_0021_01_000077/__app__.jar > /var/log/hadoop-yarn/containers/application_1572981097605_0021/container_1572981097605_0021_01_000077/stdout 2> /var/log/hadoop-yarn/containers/application_1572981097605_0021/container_1572981097605_0021_01_000077/stderr at org.apache.hadoop.util.Shell.runCommand(Shell.java:972)
> at org.apache.hadoop.util.Shell.run(Shell.java:869)
> at org.apache.hadoop.util.Shell$ShellCommandExecutor.execute(Shell.java:1170)
> at org.apache.hadoop.yarn.server.nodemanager.DefaultContainerExecutor.launchContainer(DefaultContainerExecutor.java:235)
> at org.apache.hadoop.yarn.server.nodemanager.containermanager.launcher.ContainerLaunch.call(ContainerLaunch.java:299)
> at org.apache.hadoop.yarn.server.nodemanager.containermanager.launcher.ContainerLaunch.call(ContainerLaunch.java:83)
> at java.util.concurrent.FutureTask.run(FutureTask.java:266)
> at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
> at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
> at java.lang.Thread.run(Thread.java:748)
> Container exited with a non-zero exit code 134{noformat}
> From the *stdout* logs of the exiting container we see:
> {noformat}
> #
> # A fatal error has been detected by the Java Runtime Environment:
> #
> # SIGSEGV (0xb) at pc=0x00007f825e3b0e92, pid=12611, tid=0x00007f822b5fb700
> #
> # JRE version: OpenJDK Runtime Environment (8.0_222-b10) (build 1.8.0_222-b10)
> # Java VM: OpenJDK 64-Bit Server VM (25.222-b10 mixed mode linux-amd64 compressed oops)
> # Problematic frame:
> # V [libjvm.so+0xa9ae92]
> #
> # Failed to write core dump. Core dumps have been disabled. To enable core dumping, try "ulimit -c unlimited" before starting Java again
> #
> # An error report file with more information is saved as:
> # /mnt1/yarn/usercache/hadoop/appcache/application_1572981097605_0021/container_1572981097605_0021_01_000077/hs_err_pid12611.log
> #
> # If you would like to submit a bug report, please visit:
> # http://bugreport.java.com/bugreport/crash.jsp
> #{noformat}
> Also, I am unable to enable *core dump* even though *ulimit -c* is set to *unlimited*. Can you help on how to go about this issue, and also how to get the *core dump* ?
> Steps to reproduce the issue:
> * Upload the attached parquet data file to S3 *s3://<bucket>/tables/spark_29767_parquet_table/inserted_at=201910/*
> * Create a partitioned hive table
> {code:java}
> CREATE EXTERNAL TABLE `spark_29767_parquet_table`(
> `hour` bigint,
> `title` string,
> `__deleted` string,
> `status` string,
> `transformationid` string,
> `roomid` string,
> `day` bigint,
> `notes` string,
> `nunitsfromaudit` bigint,
> `ts_ms` bigint,
> `liability` string,
> `_class` string,
> `month` bigint,
> `updatedate` struct<`date`:bigint>,
> `_id` struct<oid:string>,
> `year` bigint,
> `item` struct<name:string,brandname:string,perunitpricefromaudit:struct<currency:string,amount:string>,actualPerUnitPrice:struct<currency:string,amount:string>,category:string,itemType:string,roomAmenityId:bigint>,
> `createddate` struct<`date`:bigint>,
> `actualunits` bigint,
> `description` string)
> PARTITIONED BY (
> `inserted_at` string)
> ROW FORMAT SERDE
> 'org.apache.hadoop.hive.ql.io.parquet.serde.ParquetHiveSerDe'
> STORED AS INPUTFORMAT
> 'org.apache.hadoop.hive.ql.io.parquet.MapredParquetInputFormat'
> OUTPUTFORMAT
> 'org.apache.hadoop.hive.ql.io.parquet.MapredParquetOutputFormat'
> LOCATION
> 's3://<bucket>/tables/spark_29767_parquet_table'
> {code}
> * Sync partition
> {code:java}
> ALTER TABLE spark_29767_parquet_table ADD PARTITION (inserted_at='201910') location 's3://<bucket>/tables/spark_29767_parquet_table/inserted_at=201910/'
> {code}
> * In pyspark run the following:
> {code:java}
> // Read the base data frame
> from pyspark import SparkContext, SparkConf
> from pyspark.sql import SparkSession, HiveContext
> from pyspark.sql.functions import lit
> sparkSession = (SparkSession
> .builder
> .appName('example-pyspark-read-and-write-from-hive')
> .enableHiveSupport()
> .getOrCreate())base_df = sparkSession.table("spark_29767_parquet_table")
> base_df = sparkSession.table("spark_29767_parquet_table")
> base_df = base_df.select("_id", "_class", "roomid", "item", "inserted_at")
> // Create a new dataframe with one row for union
> from pyspark.sql import *
> import pyspark.sql.types
> from pyspark.sql.types import *
> schema = StructType([
> StructField("_id",StructType([StructField("oid",StringType(),True)]),True),
> StructField("_class",StringType(),True),
> StructField("roomid",StringType(),True),
> StructField("item",StructType([
> StructField("name",StringType(),True),
> StructField("brandname",StringType(),True),
> StructField("perunitpricefromaudit",
> StructType([
> StructField("currency",StringType(),True),
> StructField("amount",StringType(),True)]),True),
> StructField("actualperunitprice",StructType([
> StructField("currency",StringType(),True),
> StructField("amount",StringType(),True)]),True),
> StructField("category",StringType(),True),
> StructField("itemtype",StringType(),True),
> StructField("roomamenityid",LongType(),True)]),True),
> StructField("inserted_at",StringType(),True)])
> data = [
> Row(Row("5daff5ca43b8a36756c23b0f"),
> "com.oyo.transformations.tasks.model.implementations.AuditItemTaskImpl",
> None,
> Row("Geyser Installation(with accessories)",None,Row("INR", "425.0"),None,"INFRASTRUCTURE","PMC",None),
> "201910"
> )
> ]
> inc_df = spark.createDataFrame(
> spark.sparkContext.parallelize(data),
> schema
> )
> inc_df.union(base_df).show()
> {code}
>
--
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