Hadoop cluster solutions to large amounts of data hadoop command execution error or GC can not return results

As the business continues to grow, the amount of data stored in the cluster hadoop also increasing the amount of data will appear in the PB level hadoop / hdfs command performs slowly or does not return a result and GC errors;

You need to adjust the size parameters HADOOP_HEAPSIZE; 1000MB the default parameter;

The actual situation:

Data amount 1PB, HADOOP_HEAPSIZE parameter is set to 16GB, the results can be returned to normal.

Adjusted according to actual situation and ensure hadoop / hdfs normal command such as return results.

Guess you like

Origin www.cnblogs.com/songyuejie/p/11221609.html