执行”spark-shell –master yarn –deploy-mode client”,虚拟内存大小溢出，报错

在Hadoop 2.7.2集群下执行如下命令：

spark-shell  --master yarn --deploy-mode client

爆出下面的错误：

org.apache.spark.SparkException: Yarn application has already ended! It might have been killed or unable to launch application master.

在Yarn WebUI上面查看启动的Cluster状态，log显示为：

Container [pid=28920,containerID=container_1389136889967_0001_01_000121] is running beyond virtual memory limits. Current
usage: 1.2 GB of 1 GB physical memory used; 2.2 GB of 2.1 GB virtual memory used. Killing container.

这是由于虚拟内存大小超过了设定的数值，可以修改配置，进行规避。

There is a check placed at Yarn level for Vertual and Physical memory usage ratio. Issue is not only that VM doesn't have sufficient pysical memory. But it is because Virtual memory usage is more than expected for given physical memory.

Note : This is happening on Centos/RHEL 6 due to its aggressive allocation of virtual memory.

It can be resolved either by :

1. Disable virtual memory usage check by setting yarn.nodemanager.vmem-check-enabled to false;
2. Increase VM:PM ratio by setting yarn.nodemanager.vmem-pmem-ratio to some higher value（default value is 2.1）.

Add following property in yarn-site.xml
     <property>
              <name>yarn.nodemanager.vmem-check-enabled</name>
              <value>false</value>
              <description>Whether virtual memory limits will be enforced for containers</description>
    </property>
              <property>
              <name>yarn.nodemanager.vmem-pmem-ratio</name>
              <value>4</value>
              <description>Ratio between virtual memory to physical memory when setting memory limits for containers</description>
    </property>

3.Then, restart yarn.

Reference:

http://blog.cloudera.com/blog/2014/04/apache-hadoop-yarn-avoiding-6-time-consuming-gotchas/

http://blog.chinaunix.net/uid-28311809-id-4383551.html

http://stackoverflow.com/questions/21005643/container-is-running-beyond-memory-limits