Presto 0.141t Documentation

4.2. Presto VM Sandbox on HDP

« 4.1. Presto VM Sandbox on CDH 4.3. Airpal »

4.2. Presto VM Sandbox on HDP

This is a VirtualBox sandbox image that can be used to experiment with the Presto distribution. The credentials for the VM are:

user: presto
password: presto

The image is based on CentOS 6.7 and contains the following components:

  • Presto is installed in /usr/lib/presto. It is set up with Hive, TPCH and JMX connectors. Logs are located in /var/log/presto.
  • presto-admin is installed in /opt/prestoadmin. You can run presto-admin by executing: /opt/prestoadmin/presto-admin.
  • For documentation of the Presto release embedded in the VM see: Presto Documentation.

In addition, the image contains:

  • The Hortonworks Hadoop distribution, version 2.3, running in pseudo distributed mode.
  • Hive (set up to use YARN).
  • Hive Metastore.
  • Zookeeper (used by Hive).
  • MySQL (used by Hive Metastore).

The following sample tables are loaded to HDFS and are visible from Hive:

  • TPC-H nation table.
  • TPC-H region table.

The presto-cli executable JAR can be found in /home/presto and should be used to execute Presto queries. Please wait a few moments after the image boots for the Presto service to start.

Usage example:

[presto@presto-demo-hdp ~]# java -jar /home/presto/presto-cli.jar --catalog hive --schema default
show tables;
 Table
--------
 nation
 region
(2 rows)

presto:default> select * from nation;
...

JDBC Driver

You can also find the Presto JDBC JAR on the VM (/home/presto/presto-jdbc.jar). It can be used to connect to the Presto server running in the VM from your Java application.

Hadoop Services Startup

After the VM boots, some Hadoop services may still not be started. Starting them takes a couple of minutes depending on the host machine. You can monitor the startup progress of these services through the Ambari UI, which is accessible on port 8081 at http://[guest-ip-address]:8081/. In order to access Ambari from the host machine, you will need to change the network adaptor for the VM from NAT to the Bridged Adaptor. The credentials are admin/admin.

VM Networking

By default, the VM is configured to use NAT networking and will have an IP of 10.0.2.15 in VirtualBox. For some systems with local firewalls or connected to VPN this may result in being unable to reach the Airpal, presto and various hadoop web UI pages.

To resolve this, enable port forwarding via the VirtualBox UI by navigating to your VM -> settings -> Network -> Advanced. Click on the ‘Port Forwarding’ button and configure the ports you would like to forward from your local machine to the VM. For example, 127.0.0.1:50070 -> 10.0.2.15:50070 will enable access to the hadoop name node web UI.

« 4.1. Presto VM Sandbox on CDH 4.3. Airpal »