安裝與Spark相關的其他組件的時候，例如JDK，Hadoop，Yarn，Hive，Kafka等，要考慮到這些組件和Spark的版本兼容關系。這個對應關系可以在Spark源代碼的pom.xml文件中查看。
一、 下載Spark源代碼
打開網址https://github.com/apache/spark，例如選擇v2.4.0-rc5版本，再點擊“Clone or download”按鈕，點擊下方的“Download ZIP”進行下載。
二、查看pom.xml文件
將下載的源代碼壓縮包解壓后，打開里面的pom.xml文件，查看properties標簽內各配置項，里面有列出其他組件的兼容版本信息，例如<hadoop.version>2.6.5</hadoop.version>表示hadoop版本為2.6.5。如下：
<properties>
<project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
<project.reporting.outputEncoding>UTF-8</project.reporting.outputEncoding>
<java.version>1.8</java.version>
<maven.compiler.source>${java.version}</maven.compiler.source>
<maven.compiler.target>${java.version}</maven.compiler.target>
<maven.version>3.5.4</maven.version>
<sbt.project.name>spark</sbt.project.name>
<slf4j.version>1.7.16</slf4j.version>
<log4j.version>1.2.17</log4j.version>
<hadoop.version>2.6.5</hadoop.version>
<protobuf.version>2.5.0</protobuf.version>
<yarn.version>${hadoop.version}</yarn.version>
<flume.version>1.6.0</flume.version>
<zookeeper.version>3.4.6</zookeeper.version>
<curator.version>2.6.0</curator.version>
<hive.group>org.spark-project.hive</hive.group>
<!-- Version used in Maven Hive dependency -->
<hive.version>1.2.1.spark2</hive.version>
<!-- Version used for internal directory structure -->
<hive.version.short>1.2.1</hive.version.short>
<derby.version>10.12.1.1</derby.version>
<parquet.version>1.10.0</parquet.version>
<orc.version>1.5.2</orc.version>
<orc.classifier>nohive</orc.classifier>
<hive.parquet.version>1.6.0</hive.parquet.version>
<jetty.version>9.3.24.v20180605</jetty.version>
<javaxservlet.version>3.1.0</javaxservlet.version>
<chill.version>0.9.3</chill.version>
<ivy.version>2.4.0</ivy.version>
<oro.version>2.0.8</oro.version>
<codahale.metrics.version>3.1.5</codahale.metrics.version>
<avro.version>1.8.2</avro.version>
<avro.mapred.classifier>hadoop2</avro.mapred.classifier>
<aws.kinesis.client.version>1.8.10</aws.kinesis.client.version>
<!-- Should be consistent with Kinesis client dependency -->
<aws.java.sdk.version>1.11.271</aws.java.sdk.version>
<!-- the producer is used in tests -->
<aws.kinesis.producer.version>0.12.8</aws.kinesis.producer.version>
<!-- org.apache.httpcomponents/httpclient-->
<commons.httpclient.version>4.5.6</commons.httpclient.version>
<commons.httpcore.version>4.4.10</commons.httpcore.version>
<!-- commons-httpclient/commons-httpclient-->
<httpclient.classic.version>3.1</httpclient.classic.version>
<commons.math3.version>3.4.1</commons.math3.version>
<!-- managed up from 3.2.1 for SPARK-11652 -->
<commons.collections.version>3.2.2</commons.collections.version>
<scala.version>2.11.12</scala.version>
<scala.binary.version>2.11</scala.binary.version>
<codehaus.jackson.version>1.9.13</codehaus.jackson.version>
<fasterxml.jackson.version>2.6.7</fasterxml.jackson.version>
<fasterxml.jackson.databind.version>2.6.7.1</fasterxml.jackson.databind.version>
<snappy.version>1.1.7.1</snappy.version>
<netlib.java.version>1.1.2</netlib.java.version>
<calcite.version>1.2.0-incubating</calcite.version>
<commons-codec.version>1.10</commons-codec.version>
<commons-io.version>2.4</commons-io.version>
<!-- org.apache.commons/commons-lang/-->
<commons-lang2.version>2.6</commons-lang2.version>
<!-- org.apache.commons/commons-lang3/-->
<commons-lang3.version>3.5</commons-lang3.version>
<datanucleus-core.version>3.2.10</datanucleus-core.version>
<janino.version>3.0.9</janino.version>
<jersey.version>2.22.2</jersey.version>
<joda.version>2.9.3</joda.version>
<jodd.version>3.5.2</jodd.version>
<jsr305.version>1.3.9</jsr305.version>
<libthrift.version>0.9.3</libthrift.version>
<antlr4.version>4.7</antlr4.version>
<jpam.version>1.1</jpam.version>
<selenium.version>2.52.0</selenium.version>
<!--
Managed up from older version from Avro; sync with jackson-module-paranamer dependency version
-->
<paranamer.version>2.8</paranamer.version>
<maven-antrun.version>1.8</maven-antrun.version>
<commons-crypto.version>1.0.0</commons-crypto.version>
<!--
If you are changing Arrow version specification, please check ./python/pyspark/sql/utils.py,
./python/run-tests.py and ./python/setup.py too.
-->
<arrow.version>0.10.0</arrow.version>
<test.java.home>${java.home}</test.java.home>
<test.exclude.tags></test.exclude.tags>
<test.include.tags></test.include.tags>
<!-- Package to use when relocating shaded classes. -->
<spark.shade.packageName>org.spark_project</spark.shade.packageName>
<!-- Modules that copy jars to the build directory should do so under this location. -->
<jars.target.dir>${project.build.directory}/scala-${scala.binary.version}/jars</jars.target.dir>
<!-- Allow modules to enable / disable certain build plugins easily. -->
<build.testJarPhase>prepare-package</build.testJarPhase>
<build.copyDependenciesPhase>none</build.copyDependenciesPhase>
<!--
Dependency scopes that can be overridden by enabling certain profiles. These profiles are
declared in the projects that build assemblies.
For other projects the scope should remain as "compile", otherwise they are not available
during compilation if the dependency is transivite (e.g. "graphx/" depending on "core/" and
needing Hadoop classes in the classpath to compile).
-->
<flume.deps.scope>compile</flume.deps.scope>
<hadoop.deps.scope>compile</hadoop.deps.scope>
<hive.deps.scope>compile</hive.deps.scope>
<orc.deps.scope>compile</orc.deps.scope>
<parquet.deps.scope>compile</parquet.deps.scope>
<parquet.test.deps.scope>test</parquet.test.deps.scope>
<!--
Overridable test home. So that you can call individual pom files directly without
things breaking.
-->
<spark.test.home>${session.executionRootDirectory}</spark.test.home>
<CodeCacheSize>512m</CodeCacheSize>
</properties>
完畢。