@Alexander Carreño Nice to know you recreated the ambari databases and successfully launched the cluster installation. Depending on your cluster user production,test or trainging,its recommended you learn the good practice from day one. Never use embedded database but create custom database I have a preference for MariaDB or Mysql easy management and open source. You will have to create in advance databases for hive,oozie and ranger but its optional as you could do it during the configuration of the components. The disk partitioning should follow the same logic of in-built HDFS of storing 3 copies,the default replication factor of file blocks distribution in DataNodes. Having said that a good practice is to have 3 separate disk partition and different mount points. I see you are using ubuntu, you should have used custom partition eg mkdir -p /grid/0 mkdir -p /grid/1 mkdir -p /grid/2 mkdir -p /grid/x or another classic layout mkdir /u01 mkdir /u02 mkdir /u03 mkdir /u0x Please see this official cluster planning recommendation Taking into account that you have no component installed I would advice you to start afresh with a good structure. Re-install ubuntu partiion the disk accordingly see above document 3 partition of 300GB could do. Prepare the environment, firewall,ntp,passwordless etc Install MariaDB/Mysql Create the ambari database required Create databases for oozie,hive and ranger optional can be setup later. Chose the correct partition /grid/0 --/grid/x or /u01--/u0x The installation should be successful Please let me know if you need any info ... View more