Revision history [back]

click to hide/show revision 1
initial version

First of all, which process failed with OutOfMemoryError? I would say that it is not likely something controlled by client opts.

Second, and HADOOP_CLIENT_OPTS control the same params, but in different ways. Overwriting of will lead to new value in mapred-site.xml and I believe you have right value because of "I have modified". This has nothing to do with file.

Third, Sahara doesn't expect you do any job executions until cluster is in Active state. Cluster in "Waiting" state means that Sahara didn't finish hadoop installation and cluster is not ready for job executions yet. You need to fix this issue first.