I have to change the namenode heapsize (i.e. HADOOP_NAMENODE_OPTS) on a system where the web GUI is not available.
I can access the API via curl, and I can use settings.sh.
I looked at all settings using settings.sh in "hdfs-site", but there was nothing that looked remotely like what I need.
So I believe I need to talk to the API directly. Any hints, especially pointers to documentation, are greatly appreciated.
I believe the "Configuring NameNode Heap Size" manual instructions doc page at http://docs.hortonworks.com/HDPDocuments/HDP2/HDP-2.3.6/bk_installing_manually_book/content/ref-8095... can help you with this.
On the Ambari-server node:
cd /var/lib/ambari-server/resources/scripts ./configs.sh -u admin -p admin get localhost Sandbox hadoop-env > /tmp/hadoop-env.json
You can edit this file and then use the same command with a SET value to push the config back.