Apache
/
druid
mirror of https://github.com/apache/druid.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Fix Index hadoop failing with index.zip is not a valid DFS filename (#11316) 

* * Fix bug

* * simplify class loading

* * fix example configs for integration tests

* Small classloader cleanup

Co-authored-by: jon-wei <jon.wei@imply.io>
This commit is contained in:
zachjsh 2021-06-01 19:14:50 -04:00 committed by GitHub
parent a24817d20a
commit 27f1b6cbf3
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
4 changed files with 16 additions and 27 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
indexing-hadoop/src/main/java/org/apache/druid/indexer/JobHelper.java
View File

@ -476,8 +476,8 @@ public class JobHelper
return new DataSegmentAndIndexZipFilePath(
finalSegment,
tmpPath.toUri().getPath(),
finalIndexZipFilePath.toUri().getPath()
tmpPath.toUri().toString(),
finalIndexZipFilePath.toUri().toString()
);
}

27
indexing-service/src/main/java/org/apache/druid/indexing/common/task/HadoopIndexTask.java
View File

@ -450,16 +450,11 @@ public class HadoopIndexTask extends HadoopTask implements ChatHandler
List<DataSegmentAndIndexZipFilePath> dataSegmentAndIndexZipFilePaths = buildSegmentsStatus.getDataSegmentAndIndexZipFilePaths();
if (dataSegmentAndIndexZipFilePaths != null) {
indexGeneratorJobSuccess = true;
try {
Thread.currentThread().setContextClassLoader(oldLoader);
renameSegmentIndexFilesJob(
toolbox.getJsonMapper().writeValueAsString(indexerSchema),
toolbox.getJsonMapper().writeValueAsString(dataSegmentAndIndexZipFilePaths)
);
}
finally {
Thread.currentThread().setContextClassLoader(loader);
}
ArrayList<DataSegment> segments = new ArrayList<>(dataSegmentAndIndexZipFilePaths.stream()
.map(
DataSegmentAndIndexZipFilePath::getSegment)
@ -545,22 +540,20 @@ public class HadoopIndexTask extends HadoopTask implements ChatHandler
}
}
/**
* Must be called only when the hadoopy classloader is the current classloader
*/
private void renameSegmentIndexFilesJob(
String hadoopIngestionSpecStr,
String dataSegmentAndIndexZipFilePathListStr
)
{
final ClassLoader oldLoader = Thread.currentThread().getContextClassLoader();
final ClassLoader loader = Thread.currentThread().getContextClassLoader();
try {
ClassLoader loader = HadoopTask.buildClassLoader(
getHadoopDependencyCoordinates(),
taskConfig.getDefaultHadoopCoordinates()
);
Object renameSegmentIndexFilesRunner = getForeignClassloaderObject(
"org.apache.druid.indexing.common.task.HadoopIndexTask$HadoopRenameSegmentIndexFilesRunner",
loader
final Class<?> clazz = loader.loadClass(
"org.apache.druid.indexing.common.task.HadoopIndexTask$HadoopRenameSegmentIndexFilesRunner"
);
Object renameSegmentIndexFilesRunner = clazz.newInstance();
String[] renameSegmentIndexFilesJobInput = new String[]{
hadoopIngestionSpecStr,
@ -573,7 +566,6 @@ public class HadoopIndexTask extends HadoopTask implements ChatHandler
renameSegmentIndexFilesJobInput.getClass()
);
Thread.currentThread().setContextClassLoader(loader);
renameSegmentIndexFiles.invoke(
renameSegmentIndexFilesRunner,
new Object[]{renameSegmentIndexFilesJobInput}
@ -582,9 +574,6 @@ public class HadoopIndexTask extends HadoopTask implements ChatHandler
catch (Exception e) {
throw new RuntimeException(e);
}
finally {
Thread.currentThread().setContextClassLoader(oldLoader);
}
}
private void indexerGeneratorCleanupJob(

2
integration-tests/docker/environment-configs/override-examples/hadoop/s3_to_hdfs
View File

@ -31,4 +31,4 @@ AWS_REGION=<OVERRIDE_THIS>
druid_extensions_loadList=["druid-s3-extensions","druid-hdfs-storage"]
druid.indexer.task.defaultHadoopCoordinates=["org.apache.hadoop:hadoop-client:2.8.5", "org.apache.hadoop:hadoop-aws:2.8.5"]
druid_indexer_task_defaultHadoopCoordinates=["org.apache.hadoop:hadoop-client:2.8.5", "org.apache.hadoop:hadoop-aws:2.8.5"]

2
integration-tests/docker/environment-configs/override-examples/hadoop/s3_to_s3
View File

@ -32,4 +32,4 @@ AWS_REGION=<OVERRIDE_THIS>
druid_extensions_loadList=["druid-s3-extensions","druid-hdfs-storage"]
druid.indexer.task.defaultHadoopCoordinates=["org.apache.hadoop:hadoop-client:2.8.5", "org.apache.hadoop:hadoop-aws:2.8.5"]
druid_indexer_task_defaultHadoopCoordinates=["org.apache.hadoop:hadoop-client:2.8.5", "org.apache.hadoop:hadoop-aws:2.8.5"]