Avery Ching wrote: > It shouldn't be, your code looks very similar to the unittests (i.e. > TestManualCheckpoint.java). So, you're trying to run your test with the > local hadoop (similar to the unittests)? Or are you using an actual > hadoop setup?
Hi Avery, here is a few more details on what I am trying to do, in order to run my Giraph jobs on a local Hadoop running (for testing and debugging stuff locally): GiraphJob job = new GiraphJob("shortest paths"); Configuration conf = job.getConfiguration(); conf.setBoolean(GiraphJob.SPLIT_MASTER_WORKER, false); conf.setBoolean(GiraphJob.LOCAL_TEST_MODE, true); // conf.set(GiraphJob.ZOOKEEPER_JAR, "file://target/dependency/zookeeper-3.3.3.jar"); job.setWorkerConfiguration(1, 1, 100.0f); job.setVertexClass(SimpleShortestPathsVertex.class); job.setVertexInputFormatClass(SimpleShortestPathsVertexInputFormat.class); job.setVertexOutputFormatClass(SimpleShortestPathsVertexOutputFormat.class); FileInputFormat.addInputPath(job.getInternalJob(), new Path("src/main/resources/giraph1.txt")); Path outputPath = new Path("target/giraph1"); FileSystem hdfs = FileSystem.get(conf); hdfs.delete(outputPath, true); FileOutputFormat.setOutputPath(job.getInternalJob(), outputPath); job.run(true); Am I doing something wrong/stupid here? Am I missing something important? (probably! but I do not see what I am missing) This is what I think happens... In GraphMapper something goes wrong during setup(context), probably because GiraphJob.ZOOKEEPER_JAR is not set(?) and an exception different from IOException is thrown and I do not see any useful error message: try { setup(context); while (context.nextKeyValue()) { map(context.getCurrentKey(), context.getCurrentValue(), context); } cleanup(context); } catch (IOException e) { if (mapFunctions == MapFunctions.WORKER_ONLY) { serviceWorker.failureCleanup(); } throw new IllegalStateException( "run: Caught an unrecoverable exception " + e.getMessage(), e); } My question is: is it possible to run a Giraph job as I am trying to do above (for testing only) or developers need to have an Hadoop cluster either remote or locally and ZooKeeper running (either remote or locally)? Thanks, Paolo