Nodes

Using the Nodes page

The Nodes section provides a comprehensive view of all of the nodes that are used across your cluster. There are two tabs for this screen, Agents and Masters. By default, you will see the Agents tab.

Figure 1 - Agents tab

Agents tab

This is the default view. Also by default, all of your nodes are displayed in List view, sorted by health. You can filter this list by framework or service name.

Figure 2 - Filter by framework

If you prefer a more graphical display, click on the icons on the right side of the screen to switch between a list and “donuts” view of the nodes.

Figure 3 - Nodes “donut” view

Masters tab

The Masters tab displays information about the masters in the cluster. You can see the leader and non-leaders in the cluster, with their corresponding IP and port, region, version, started time, and elected time.

Figure 4 - Masters tab

Nodes instance

Clicking on a Host IP opens a Nodes instance page, which provides more information about the node. You can use the Search box or the drop down filter to narrow your list. The Nodes instance page has three tabs: Tasks, Health, and Details. The Tasks tab is displayed by default.

Tasks

Figure 5 - Tasks tab

The Tasks tab displays the following information:

Name	Description
ID	Unique identifier for this instance of a service. Consists of the name of the service (for example, `confluent-kafka` plus configuration details.)
Name	Unique name of this instance of a service. Consists of a series of words separated by dashes (for example, `data-science-engine`). Each word must be at least 1 alphanumeric character and may only contain digits (`0-9`), dashes (`-`), dots (`.`), and lowercase letters (`a-z`). The word may not begin or end with a dash.
Host	IP address of this node.
Zone	Cloud provider zone.
Region	Cloud provider region. If left undefined this will run in your local region.
Status	The current state of the node.
Health	The current health of the node.
Logs	Clicking on the “View Logs” icon will open the logs for this instance of the service.
CPU	The number of CPU shares allocated to the process.
Mem	The amount of memory, in MB, allocated to the process.
GPU	Number of GPUs available to this node.
Updated	Time, in minutes, since last update.

A vertical line of dots in the upper right corner opens a menu which allows you to drain or deactivate a node.

Figure 6 - Deactivation menu

For more information on draining nodes, see Draining a Node. For more information on deactivating a node, see the documentation on how to Shut Down and Decomission Nodes.

Health

The Health tab displays the status of your node’s health checks.

Figure 7 - Nodes > Health tab

You can filter this list using the All Health Checks drop down menu.

Figure 8 - Health checks filter

Details

The Nodes > Details tab will show you the identification information for this node, as well as its Status and Resources.

Figure 9 - Nodes details tab

Service details

If you click on the name of the service (for example, HDFS), you will see the Nodes > Service Details page. This will display information about the configuration, labels, Marathon configuration, and health check status.

Figure 10 - Service details

Files tab

The Files tab for your service displays a Working Directory, which will display the Name, Permissions, Owner, Size and last Modification Date for this cluster.

Figure 11 - Files tab

Logs tab

If you click on the Logs tab, you will see the log kept by DC/OS for your cluster.

Figure 12 - Logs

You can filter the results by stderr and stdout filters. To download a copy of the log, click on the down arrow on the right hand side of the menu.

For complete details on logging, see the Logging documentation. You can also find the command line interface commands for logging in the CLI Command Reference documentation.

Host details

If you click on the IP address of the host, you will see the Tasks view by default.

Figure 13 - Node tasks

CLI equivalent commands

For information about the command line interface commands for Nodes, see the CLI Command Reference documentation.