CDH5.2: MR, Unable to initialize any output collector

问题

Cloudera CDH5.2 Quickstart VM Cloudera Manager showing all nodes state = GREEN

I've jared on Eclipse a MR job including all relevant cloudera jars in the Build Path: avro-1.7.6-cdh5.2.0.jar, avro-mapred-1.7.6-cdh5.2.0-hadoop2.jar, hadoop-common-2.5.0-cdh5.2.0.jar, hadoop-mapreduce-client-core-2.5.0-cdh5.2.0.jar

I've run the following job

hadoop jar jproject1.jar avro00.AvroUserPrefCount -libjars ${LIBJARS} avro/00/in avro/00/out

I get the following error, is it a Java heap problem, any comments ? Thank you in advance

14/11/14 01:02:40 INFO client.RMProxy: Connecting to ResourceManager at quickstart.cloudera/127.0.0.1:8032
14/11/14 01:02:43 INFO input.FileInputFormat: Total input paths to process : 1
14/11/14 01:02:43 INFO mapreduce.JobSubmitter: number of splits:1
14/11/14 01:02:44 INFO mapreduce.JobSubmitter: Submitting tokens for job: job_1415950730849_0001
14/11/14 01:02:45 INFO impl.YarnClientImpl: Submitted application application_1415950730849_0001
14/11/14 01:02:45 INFO mapreduce.Job: The url to track the job: http://quickstart.cloudera:8088/proxy/application_1415950730849_0001/
14/11/14 01:02:45 INFO mapreduce.Job: Running job: job_1415950730849_0001
14/11/14 01:03:04 INFO mapreduce.Job: Job job_1415950730849_0001 running in uber mode : false
14/11/14 01:03:04 INFO mapreduce.Job:  map 0% reduce 0%
14/11/14 01:03:11 INFO mapreduce.Job: Task Id : attempt_1415950730849_0001_m_000000_0, Status : FAILED
Error: java.io.IOException: Unable to initialize any output collector
    at org.apache.hadoop.mapred.MapTask.createSortingCollector(MapTask.java:412)
    at org.apache.hadoop.mapred.MapTask.access$100(MapTask.java:81)
    at org.apache.hadoop.mapred.MapTask$NewOutputCollector.<init>(MapTask.java:695)
    at org.apache.hadoop.mapred.MapTask.runNewMapper(MapTask.java:767)
    at org.apache.hadoop.mapred.MapTask.run(MapTask.java:341)
    at org.apache.hadoop.mapred.YarnChild$2.run(YarnChild.java:168)
    at java.security.AccessController.doPrivileged(Native Method)
    at javax.security.auth.Subject.doAs(Subject.java:415)
    at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1614)
    at org.apache.hadoop.mapred.YarnChild.main(YarnChild.java:163)
...
...

回答1:

Checking the full task log of the failed attempt attempt_1415950730849_0001_m_000000_0 will help tell why you ran into the given exception.

The most common reason of observing such an error is a misconfigured value of io.sort.mb in your job. Its value must never be anywhere close to (or higher than) the configured map task heap size, and must also not currently exceed ~2000 MB (Java array maximum size).

An upstream improvement of making the error more clear on the true failure was also filed and resolved recently, via MAPREDUCE-6194.

回答2:

I encountered the same issue yesterday. I checked the syslog for the particular map task which was failing, which suggested that I was getting another exception in that task which was triggering this error. In my case this was an invalid parsing, and when I corrected that issue, this error was fixed.

Closer examination of the log for the failed task should give you the root cause for the issue.

来源：https://stackoverflow.com/questions/26926571/cdh5-2-mr-unable-to-initialize-any-output-collector