[ https://issues.apache.org/jira/browse/ACCUMULO-4579?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15847705#comment-15847705 ]
ASF GitHub Bot commented on ACCUMULO-4579: ------------------------------------------ Github user ctubbsii commented on a diff in the pull request: https://github.com/apache/accumulo-testing/pull/3#discussion_r98796795 --- Diff: core/src/main/java/org/apache/accumulo/testing/core/TestEnv.java --- @@ -96,15 +97,22 @@ public String getPid() { } public Configuration getHadoopConfiguration() { - Configuration config = new Configuration(); - config.set("mapreduce.framework.name", "yarn"); - // Setting below are required due to bundled jar breaking default - // config. - // See - // http://stackoverflow.com/questions/17265002/hadoop-no-filesystem-for-scheme-file - config.set("fs.hdfs.impl", org.apache.hadoop.hdfs.DistributedFileSystem.class.getName()); - config.set("fs.file.impl", org.apache.hadoop.fs.LocalFileSystem.class.getName()); - return config; + if (hadoopConfig == null) { + String hadoopPrefix = System.getenv("HADOOP_PREFIX"); + if (hadoopPrefix == null || hadoopPrefix.isEmpty()) { + throw new IllegalArgumentException("HADOOP_PREFIX must be sent in env"); + } + hadoopConfig = new Configuration(); + hadoopConfig.addResource(new Path(hadoopPrefix + "/etc/hadoop/core-site.xml")); --- End diff -- I think using the properties file is better. I don't think we should be writing java code to depend on env variables, especially arbitrary script conventions like `HADOOP_PREFIX`, which is very sensitive to Hadoop packaging/deployment changes. > Continous ingest failing due to bad hadoop configuration > -------------------------------------------------------- > > Key: ACCUMULO-4579 > URL: https://issues.apache.org/jira/browse/ACCUMULO-4579 > Project: Accumulo > Issue Type: Bug > Components: test > Affects Versions: 2.0.0 > Reporter: Mike Walch > Assignee: Mike Walch > Fix For: 2.0.0 > > > I ran the continuous ingest test in the accumulo-testing repo on a > distributed cluster. The test failed due to Twill storing configuration on > the local file system rather than HDFS. This is occurring because the the > YarnTwillRunnerService is not being provide with the proper hadoop > configuration. -- This message was sent by Atlassian JIRA (v6.3.15#6346)