#tasktracker

Rishab Chandra, co-founder of Task Tracker Suite, uses his expertise in SaaS and consulting to create innovative tools that enhance business productivity.

Sometimes your MR job is lenient enough to allow few task failures but the default setting of the leniency is set to fail a job on any task failures.

# of failed Map Tasks exceeded allowed limit. FailedCount: 1.

There isn't anywhere to change this setting in mapred-default.xml.

This customization actually exists as followed in [HADOOP-1144]

Add following to your mapred-site.xml

<property>    <name>mapred.max.map.failures.percent</name>    <value>10</value>    <description>Defaults to 0, meaning any map task failure results to the job failure.</description> </property>

<property>    <name>mapred.max.reduce.failures.percent</name>    <value>10</value>    <description>Defaults to 0, meaning any reduce task failure results to the job failure.</description> </property>

These configurations are also related to

JobConf.getMaxMapTaskFailuresPercent() JobConf.getMaxReduceTaskFailuresPercent()

Note. Keep in mind that a task failes after 4 retries (mapred.map.max.attempts, mapred.reduce.max.attempts).

In Apache Hive 0.9.0, the tasktracker http address will be generated with this part "/tasklog?taskid=".

But if you use it with CDH3u5 Hadoop, the address had been changed in previous version. In that version of CDH3, the address must be "/tasklog?attemptid=".

So if a failure occur in your HiveQL in that situation, Hive automatically try to show the details of the task failure as default and fail with HTTP status code 400 because the URL is wrong.

To avoid this cheap error, you have 2 measures.

1. (Easy but temporarily) set hive.exec.show.job.failure.debug.info=false

2. modify the following 2 source codes.

org.apache.hadoop.hive.ql.exec.HadoopJobExecHelper.getTaskAttemptLogUrl(String, String)