Unlike HortonWorks or Cloudera, AWS EMR does not seem to provide any graphical interface for changing the xml configurations of the various hadoop ecosystem infrastructures.
Logging into my namenode EMR and doing a quick
find \ -iname yarn-site.xml
I managed to find it in /etc/hadoop/conf.empty/yarn-site.xml
and the performance planner to be located at /etc/hadoop/conf.empty/capacity-scheduler.xml
.
But pay attention to how they are under conf.empty, and I suspect that these may not be the actual locations for the hml site and the xmls bandwidth scheduler.
I understand that I can change these configurations when creating a cluster, but I need to know how to change them without breaking the cluster.
I just want to play planning properties, etc. and try different schedulers to determine what can work with my spark applications.
Thanks in advance!
source
share