Brainspace System Health Checks

The checks below are to be done via the command line and are for helping to ensure that the Brainspace Servers are functioning appropriately.

Disk Usage and Mount Point

It is important to ensure that there is adequate space on all drives mounted to the Brainspace Servers.

On all servers you can run the df -h command to see what is mounted, where it is mounted, how much space is used, as well as how much is available for use, an example of the output from an Application Server is below:

This output will be similar between the Application Server and Analytics Server, the OnDemand Analytics will have the addition of /localdata-share as a mount point starting in 6.3 that will be similar to the following:

Checking Services

To check the services on the Brainspace Servers starting in 6.2 we use the systemctl status command instead of the service command, you can still use the service command as it has been linked to the systemctl command. If a service is found to not be running then can be started with systemctl start <service> command. The following services should be running on each server.

Application Server:

brainspace-platform

brainspace-haproxy

postgresql-9.6

Analytics and On-Demand Analytics Server:

brainspace-analysis

If any of these services are not running and return an error when starting please contact Brainspace Support and provide the output shown on the screen.

Memory Usage

Memory (RAM) on each server plays a crucial role in how Brainspace operates and the tasks done by each service (brainspace-platform &amp; brainspace-analysis). Brainspace will allocate 70% of the total amount of RAM on the server to itself for the tasks it needs to Ingest/Build a Brain/Create Focus/etc.

To check the amount of RAM being used overall you used the free -m command which will show you how much RAM is installed on the server, how much is free, used, and available. Output would be similar to:

To see how much memory Brainspace requires to run all datasets in your environment you would run the dataset-ram.sh script, more info of which can be found here.

Disk Performance

If you feel that there may be an issue with disk performance there are two quick ways to gather statics on read/writes. For a general overview you can used the iostat command and for more depth broken down into 10 minute increments you can use sar -b or sar -p -d.