[HADOOP-16309] Unable to build Hadoop natively on ARM, node.js is being downloaded for x86

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 3.3.0
• Component/s: build
• Labels: build

Environment

Linux ARM

• Assignee: Unassigned
• Reporter: AiBe Gee
• Created: 12/May/19 03:11
• Updated: 12/May/19 03:38

Description

I'm trying to build Hadoop 3.3.0 - git snapshot from 10 May 2019 natively on Raspberry Pi2B (ARM-armv7) and failing in the YARN Application Catalog Webapp component. Apparently the node.js version is downloaded for x86 instead of armv7 and it cannot be executed - Exec format Error
Snippet from the build log:
...
Downloaded from central: https://repo.maven.apache.org/maven2/org/eclipse/sisu/org.eclipse.sisu.inject/0.0.0.M2a/org.eclipse.sisu.inject-0.0.0.M2a.jar (202 kB at 39 kB/s)
Downloaded from central: https://repo.maven.apache.org/maven2/org/apache/httpcomponents/httpcore/4.4.3/httpcore-4.4.3.jar (327 kB at 52 kB/s)
Downloaded from central: https://repo.maven.apache.org/maven2/commons-codec/commons-codec/1.9/commons-codec-1.9.jar (264 kB at 34 kB/s)
[INFO] Installing node version v8.11.3
[INFO] Downloading https://nodejs.org/dist/v8.11.3/node-v8.11.3-linux-x86.tar.gz to /root/.m2/repository/com/github/eirslett/node/8.11.3/node-8.11.3-linux-x86.tar.gz
[INFO] No proxies configured
[INFO] No proxy was configured, downloading directly
[INFO] Unpacking /root/.m2/repository/com/github/eirslett/node/8.11.3/node-8.11.3-linux-x86.tar.gz into /kit/hadoop-3-3-0/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-applications/hadoop-yarn-applications-catalog/hadoop-yarn-applications-catalog-webapp/target/node/tmp
[INFO] Copying node binary from /kit/hadoop-3-3-0/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-applications/hadoop-yarn-applications-catalog/hadoop-yarn-applications-catalog-webapp/target/node/tmp/node-v8.11.3-linux-x86/bin/node to /kit/hadoop-3-3-0/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-applications/hadoop-yarn-applications-catalog/hadoop-yarn-applications-catalog-webapp/target/node/node
[INFO] Installed node locally.
[INFO] Installing Yarn version v1.7.0
[INFO] Downloading https://github.com/yarnpkg/yarn/releases/download/v1.7.0/yarn-v1.7.0.tar.gz to /root/.m2/repository/com/github/eirslett/yarn/1.7.0/yarn-1.7.0.tar.gz
[INFO] No proxies configured
[INFO] No proxy was configured, downloading directly
May 10, 2019 10:48:32 AM org.apache.http.client.protocol.ResponseProcessCookies processCookies
WARNING: Invalid cookie header: "Set-Cookie: has_recent_activity=1; path=/; expires=Fri, 10 May 2019 08:48:32 -0000". Invalid 'expires' attribute: Fri, 10 May 2019 08:48:32 -0000
May 10, 2019 10:48:32 AM org.apache.http.client.protocol.ResponseProcessCookies processCookies
WARNING: Invalid cookie header: "Set-Cookie: logged_in=no; domain=.github.com; path=/; expires=Tue, 10 May 2039 07:48:32 -0000; secure; HttpOnly". Invalid 'expires' attribute: Tue, 10 May 2039 07:48:32 -0000
[INFO] Unpacking /root/.m2/repository/com/github/eirslett/yarn/1.7.0/yarn-1.7.0.tar.gz into /kit/hadoop-3-3-0/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-applications/hadoop-yarn-applications-catalog/hadoop-yarn-applications-catalog-webapp/target/node/yarn
[INFO] Installed Yarn locally.
[INFO]
[INFO] — frontend-maven-plugin:1.6:yarn (yarn install) @ hadoop-yarn-applications-catalog-webapp —
[INFO] Running 'yarn ' in /kit/hadoop-3-3-0/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-applications/hadoop-yarn-applications-catalog/hadoop-yarn-applications-catalog-webapp/target
[ERROR] /kit/hadoop-3-3-0/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-applications/hadoop-yarn-applications-catalog/hadoop-yarn-applications-catalog-webapp/target/node/yarn/dist/bin/yarn: line 18: /kit/hadoop-3-3-0/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-applications/hadoop-yarn-applications-catalog/hadoop-yarn-applications-catalog-webapp/target/node/node: cannot execute binary file: Exec format error
[ERROR] /kit/hadoop-3-3-0/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-applications/hadoop-yarn-applications-catalog/hadoop-yarn-applications-catalog-webapp/target/node/yarn/dist/bin/yarn: line 18: /kit/hadoop-3-3-0/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-applications/hadoop-yarn-applications-catalog/hadoop-yarn-applications-catalog-webapp/target/node/node: Success
[INFO] ------------------------------------------------------------------------
[INFO] Reactor Summary for Apache Hadoop Main 3.3.0-SNAPSHOT:
[INFO]
[INFO] Apache Hadoop Main ................................. SUCCESS [ 10.305 s]
[INFO] Apache Hadoop Build Tools .......................... SUCCESS [ 11.622 s]
[INFO] Apache Hadoop Project POM .......................... SUCCESS [ 11.115 s]
.....
[INFO] Apache Hadoop YARN Application Catalog ............. SUCCESS [ 1.043 s]
[INFO] Apache Hadoop YARN Application Catalog Webapp ...... FAILURE [07:48 min]
[INFO] Apache Hadoop YARN Application Catalog Docker Image SKIPPED
[INFO] Apache Hadoop YARN Application MaWo ................ SKIPPED
....
[INFO] Apache Hadoop Cloud Storage Project ................ SKIPPED
[INFO] ------------------------------------------------------------------------
[INFO] BUILD FAILURE
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 04:02 h
[INFO] Finished at: 2019-05-10T10:48:40+03:00
[INFO] ------------------------------------------------------------------------
[ERROR] Failed to execute goal com.github.eirslett:frontend-maven-plugin:1.6:yarn (yarn install) on project hadoop-yarn-applications-catalog-webapp: Failed to run task: 'yarn ' failed. org.apache.commons.exec.ExecuteException: Process exited with an error: 126 (Exit value: 126) -> [Help 1]

I was able to identify the version of node.js in:
/kit/hadoop-3-3-0/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-applications/hadoop-yarn-applications-catalog/hadoop-yarn-applications-catalog-webapp/pom.xml

<execution>
<id>install node and yarn</id>
<goals>
<goal>install-node-and-yarn</goal>
</goals>
<phase>generate-resources</phase>
<configuration>
<nodeVersion>v8.11.3</nodeVersion>
<yarnVersion>v1.7.0</yarnVersion>
</configuration>
</execution>

But failed to find where it makes the actual ARM / x86 binary selection.
It should download, unpack and execute:
https://nodejs.org/dist/v8.11.3/node-v8.11.3-linux-armv7l.tar.gz

Any help pointing me to the right file to patch/change in order to force the proper node.js version choice (ARMv7) is appreciated!

Just for reference, this is the build command:
nohup mvn package -Pdist,native,docs -DskipTests -Dtar 2>&1 | tee hadoop-build.log &

And these are my environment settings:
export PATH="$PATH:/opt/java/bin"
export M2_HOME=/opt/apache-maven-3.6.1
export "PATH=$PATH:$M2_HOME/bin"
JAVA_HOME=/opt/java
export JAVA_HOME
export ARCH=arm
export CFLAGS="-march=armv7-a -mtune=cortex-a7 -mfpu=neon-vfpv4 -mvectorize-with-neon-quad -mfloat-abi=hard"
export CXXFLAGS="-march=armv7-a -mtune=cortex-a7 -mfpu=neon-vfpv4 -mvectorize-with-neon-quad -mfloat-abi=hard"
export CPPFLAGS="-march=armv7-a -mtune=cortex-a7 -mfpu=neon-vfpv4 -mvectorize-with-neon-quad -mfloat-abi=hard"

The patches I needed to apply - otherwise the build would have failed before the reported issue:
1.
https://stackoverflow.com/questions/50661648/spring-boot-fails-to-run-maven-surefire-plugin-classnotfoundexception-org-apache/50661649#50661649
https://maven.apache.org/surefire/maven-surefire-plugin/examples/class-loading.html

• added in pom.xml:
  <plugin>
  <groupId>org.apache.maven.plugins</groupId>
  <artifactId>maven-surefire-plugin</artifactId>
  <version>3.0.0-M3</version>
  <configuration>
  <useSystemClassLoader>false</useSystemClassLoader>
  </configuration>
  </plugin>

2.
https://issues.apache.org/jira/browse/HADOOP-9320

• patch - v2.8.patch
  cd /kit/hadoop-3-3-0/hadoop-common-project/hadoop-common/
  wget https://patch-diff.githubusercontent.com/raw/apache/hadoop/pull/224.patch
  patch < 224.patch