Set Up Datameer on MapR
Installating Datameer on MapR
The Datameer software is available for installation from your my.datameer archive repository located at my.datameer.com. You must have the MapR repository configured properly in order to install Datameer. See MapR documentation for instructions.
Datameer is currently available for distribution for Mac, Redhat, Suse, Unbuntu, or Windows named mapr-zip
.
Dependencies
Datameer requires that mapr-core
is available on the host machine and that a fully configured MapR cluster is connected to the host machine. These packages must be installed and the MapR services configured prior to installing mapr-zip
.
Unless otherwise noted, all shell commands listed in this document should be run as the root user or via sudo
where available.
Pre-installation
The Datameer package built for MapR Hadoop distributions doesn't include native libs, so you need copy the libs from the cluster to Datameer using the following command:
cp -r /opt/mapr/hadoop/<hadoop_version>/lib/
native
<DATAMEER_HOME>/lib/
Reviewing the installation
Once the package has been installed successfully, the Datameer service automatically starts and is ready for use on port 8080 of the host machine. Make sure to use a Flash-enabled and supported web browser when accessing this URL.
The Datameer software is installed on the machine in /opt/datameer/datameer-3.1.1-mapr
.
ppmclala@ubuntu:/opt/datameer/das-3.1.1-mapr$ ls -l total 14812 drwxr-xr-x 3 root root 4096 2011-06-27 09:06 bin drwxr-xr-x 2 root root 4096 2011-06-27 09:06 conf drwxr-xr-x 2 root root 4096 2011-06-27 08:38 contexts drwxr-xr-x 7 root root 4096 2011-06-27 09:06 data drwxr-xr-x 5 root root 4096 2011-06-27 09:06 etc drwxr-xr-x 2 root root 4096 2011-06-27 09:06 job-jar drwxr-xr-x 13 root root 4096 2011-06-27 09:06 lib drwxr-xr-x 3 root root 4096 2011-06-27 09:06 LICENSES -rw-r--r-- 1 root root 1061 2011-06-27 08:38 LICENSES.txt drwxr-xr-x 2 root root 4096 2008-11-13 09:18 logs drwxr-xr-x 2 root root 4096 2011-06-27 09:06 plugins -rw-r--r-- 1 root root 13157439 2011-06-27 08:38 plugin-sdk-3.1.1.zip -rw-r--r-- 1 root root 17125 2008-11-13 18:09 start.jar -rw-r--r-- 1 root root 1933022 2011-06-27 08:38 stripped-dap-3.1.1-job.jar -rw-r--r-- 1 root root 3 2011-06-27 08:38 VERSION drwxr-xr-x 3 root root 4096 2011-06-27 09:06 webapps
All Datameer-related files are stored in this directory including control scripts and logs. The bin
directory contains all executable scripts, most notably conductor.sh which controls the Datameer server. You can also find SQL scripts for database initialization and various other tools. All logging output is directed to files in the logs
directory, the most interesting of which is conductor.log
. This is where the most valuable logging information related to running the Datameer system will be written.
Errors During Installation
After installing the mapr-datameer package
, if you see an error similar to:
ERROR Cidcache fs/client/fileclient/cc/cidcache.cc:1017 Thread: 1075886400 Lookup of volume mapr.cluster.root failed, error Connection reset by peer(104), CLDB: 172.16.148.131:7222 trying another CLDB
or in the Datameer conductor.log
java.io.IOException: Could not create FileClient
then you have most likely installed Datameer without first properly configuring a MapR cluster. Stop the Datameer application with the following command:
/etc/init.d/das-conductor stop
Configure and start the MapR components and then re-start the Datameer application using the command:
/etc/init.d/das-conductor start
Stopping and starting Datameer
Datameer starts automatically upon successful installation. The mapr-datameer
package also installs a script in /etc/init.d/das-conductor
which can be used to control the Datameer process.
In case Datameer doesn't start automatically (only if you have experienced errors during installation), you can run the startup script located in /etc/init.d/das-conductor
with the start option.
/etc/init.d/das-conductor start
This launches the server and starts a web application listening to http://localhost:8080/
.
To stop a running instance of Datameer, use the stop option by entering the following command:
/etc/init.d/das-conductor stop
Another supported option is shown here:
Usage: /opt/datameer/das-1.3-mapr/bin/jetty.sh {start|stop|run|restart|check|supervise} [ CONFIGS ... ]
Logging In and Using Datameer
After activating your trial license, navigate to the Login screen either by clicking Login Now on the trial activation screen or by visiting port 8080 on the machine where Datameer is installed. You can now login with the default administrator account by entering the credentials below and clicking Login.
Username | Password |
---|---|
admin | admin |
Connecting Datameer to a MapR Cluster with Zero Configuration HA for Resource Manager
MAPR Zero Configuration is automatically detected by Datameer when the Hadoop Cluster is configured with MAPR grid mode.
As of Datameer v6.1
An additional label with a checkmark is displayed on the cluster settings page when Zero Configuration HA is enabled.
Learn more about MapR's Zero Configuration Failover for the ResourceManager.
MapR Security
According to Enabling and Disabling Security Features and Getting started with MapR Security, MapR is runs cluster nodes and Java virtual machines (JVMs) with its own truststore.
That truststore is routed to:
-Djavax.net.ssl.trustStore=/opt/mapr/conf/ssl_truststore
The default password of this truststore might be set to mapr123.
If you want to know the location of the truststore that is used by MapR, use the following command:
ps -ef | grep "javax.net.ssl.trustStore"
This command shows all the Java processes which have a non-default truststore set.
To update this truststore or to trust an Amazon S3 certificate, you need the password of the ssl_truststore. Alternatively, you can overwrite the truststore to your own one.