Installation

Install GraphGrid

The first step is to download the compressed file of your edition of choice from GraphGrid.com.

It should look something like this:

graphgrid-ai-2.0.1-unix.tar.gz

You can now unzip the file by either using an unzip feature in your OS or by running:

tar -xf graphgrid-ai-2.0.1-unix.tar.gz

Next, in a terminal, move into the project directory.

cd graphgrid-ai-2.0.1

You're ready to install!

The GraphGrid install script will download all necessary dependencies. There is an optional --data-dir flag that will separate the data directory. Using the --data-dir flag is recommended for environments where data protection and management is imported in production.

Use the GraphGrid install script:

./bin/graphgrid install <opt>

note

For security purposes it is advised to change all default passwords using the password generation command. This command must be run immediately after installation. This command will not work after the start script has been run.

Start up

Start the package

./bin/graphgrid start

note

GraphGrid Curator starts up before most other services. If you navigate to the Curator dashboard immediately after start up, the Manager Model and Search may not work immediately. To check if the services are running, try the module’s status API endpoint or running docker ps -a to see a list of all running containers. The initialization time of these modules vary.

It is common to get this error on startup which means that the max amount of Docker networks are in use:

$ ./bin/graphgrid start
Starting GraphGrid
docker-compose up -d
Creating network "graphgrid" with driver "bridge"
ERROR: could not find an available, non-overlapping IPv4 address pool among the defaults to assign to the network

To fix this, use a Docker command to remove all unused networks:

docker network prune

The bin/graphgrid script also supports other commands like stop and restart. Run ./bin/graphgrid help or ./bin/graphgrid <command> help for more information.

Start and Stop Individual Services

This script also supports starting individual services. If you wish to only run a particular module, you can run ./bin/graphgrid start <module_name>.

note

You may also run the disable command to disable an entire module and its dependent services on start up. See here for details.

For example, if you only wanted to start up only the Manager and Security modules:

./bin/graphgrid start manager security

The script will only start the dependencies for each particular module you are starting.

info

If you're seeing a lot of modules "spam" the docker logs with messages like this:

file_1           | 2021-08-05 20:15:34.743 [apm-reporter] ERROR co.elastic.apm.agent.report.IntakeV2ReportingEventHandler - Failed to handle event of type METRICS with this error: apm-server
file_1           | 2021-08-05 20:15:34.743 [apm-reporter] INFO co.elastic.apm.agent.report.IntakeV2ReportingEventHandler - Backing off for 36 seconds (±10%)

It is not an error that will affect the modules running, and can be resolved by starting up the Kibana and the APM server in the packaging: ./bin/graphgrid start kibana apm-server This is likely to occur when using the packaging and only starting up a few services.

For more information about the GraphGrid scripts, watch our GraphGrid Docker and Script Tips video here.

Enable/Disable Modules and Services

This command will disable a module and its dependent services from starting up with the rest of the enabled services. The disable command should be used before start up. The install script must be run before the enable/disable commands will work.

./bin/graphgrid disable <module>

Running this will add a donotstart profile to the docker-compose.yml file.

To re-enable a module and its services, run the enable command to restore the default startup behavior.

./bin/graphgrid enable <module>

The enable command will remove the donotstart profile in docker-compose.yml.

Stop GraphGrid

To stop all running containers use the GraphGrid stop script:

./bin/graphgrid stop

note

If a user tries ./bin/graphgrid stop and it doesn’t immediately stop all the containers and gives “errors” (see picture), keep trying to rerun the stop command. This can happen when too many resources are being used and docker can’t stop everything at once. Eventually everything will be stopped if the user keeps running the stop command.