Configuration

Before starting MapD Core, the persistent data directory must be initialized. To do so, first create an empty directory at the desired path, such as /var/lib/mapd/data. Create the environment variable $MAPD_STORAGE.

Change the owner of the directory to the user that the server will run as ($MAPD_USER):

sudo mkdir -p $MAPD_STORAGE
sudo chown -R $MAPD_USER $MAPD_STORAGE

Where $MAPD_USER is the system user account that the server runs as, such as mapd, and $MAPD_STORAGE is the desired path to the MapD Core Database data directory.

Finally, run $MAPD_PATH/bin/initdb with the data directory path as the argument:

$MAPD_PATH/bin/initdb $MAPD_STORAGE

Configuration file

MapD Core supports storing options in a configuration file. This is useful if, for example, you need to run the MapD Core Server and Web Server on different ports than their defaults.

If you store a copy of mapd.conf in the $MAPD_STORAGE directory, the configuration settings are picked up automatically by the sudo systemctl start mapd_server and sudo systemctl start mapd_web_server commands.

Set the flags in the configuration file using the format <flag> = <value>. The following is a sample configuration file, with the read-only flag set to true.

port = 9091
http-port = 9090
data = "/var/lib/mapd/data"
read-only = true


[web]
port = 9092
frontend =
   "/home/osboxes/installs/mapd-3.0.0-20170502-9e5ba95-Linux-x86_64-render/frontend"

To use options provided in this file, provide the path the the config file to the --config flag of mapd_server and mapd_web_server. For example:

$MAPD_PATH/bin/mapd_server --config $MAPD_STORAGE/mapd.conf

Configuration Settings

Flag	Description	Default	Why Change It?
-h [ --help ]	Print help messages.	N/A	N/A
--config arg	Path to mapd.conf.	none	One use case might be to temporarily set a different configuration file during testing and troubleshooting.
--data arg	Directory path to MapD catalogs.	$PWD/data	You can set the path anywhere you choose.
--cpu	Set this flag on a GPU installation to instruct it to use only CPUs. (You do not have to explicitly set this flag for a CPU- only installation. It knows what it is.)	false	One use case for disabling GPUs is during database conversion, which requires moving a large amount of data with minimal processing.
--gpu	Run on GPUs and CPUs.	true	Default.
--read-only	Enable read-only mode.	false	Prevents inadvertent (or nefarious) changes to the dataset.
-p [ --port ] arg	Port number	9091	Change the port number if it collides with another service on the host. Ideally, your host only runs MapD services.
--ldap-uri arg	ldap server uri	N/A	N/A
--ldap-ou-dc arg	ldap Organizational Unit and Domain Component	=ou=users, dc=mapd,dc=com	N/A
--http-port arg	HTTP port number	9090	Change the port number if it collides with another service on the host. Ideally, your host only runs MapD services.
--flush-log	Force aggressive log file flushes.	false	When you set this the system writes messages to disk as they are generated, rather than holding them until a particular threshold is reached.
--num-gpus arg	Number of gpus to use	-1	In a shared environment, you can assign the number of GPUs to a particular application. The default is -1, which means use all available GPUs.
--start-gpu arg	First gpu to use	0	In a shared environment, if you want to reserve a set number of GPUs for a particular process, you can configure another process to use GPUs starting at a higher device ID.
-v [ --version ]	Print release version number.	N/A	N/A
--db-convert arg	Directory path to MapD database from which to convert	N/A	Set it to the path of an existing MapD <=2.04 database to be converted for MapD 2.1.0.