Hello @mm25, /opt/cloudera/parcels/CDH-5.13.1-1.cdh5.13.1.p0.2/lib/hive/conf/hive-site.xml is merely a stock file that is not used to start a service. When you manage a cluster with Cloudera Manager, Cloudera Manager will initiate the start of the service. It will use what you have configured in Cloudera Manager to generate all files necessary to start the process. This includes hive-site.xml. So, in short, the answer to why Hive is trying to use postgres is that you have it configured to do so in the configuration stored in Cloudera Manager. To affect change you will need to change the configuration within Cloudera Manager's configuration. - go to Clusters --> Hive - Click on Configuration (subtab) This will display your configuration. - On the left side, Click on Hive Metastore Database to view all database-related configuraiton options. - You can update the configuration and Save/Restart Hive to have the changes take effect. If you want to view the generated configuration files, you can view them in a Role's Processes subtab or view them at the commandline in /var/run/cloudera-scm-agent/process Each time a process is started by Cloudera Manager, a new process directory is created under that dir. I generally use ls -lrt |grep hive or something of the sort to see the most recent process directory created for a service's roles. Hope that does the trick!
... View more
