Graphical User Interface

Online Help Table of Contents

This help page details the graphical user interface for the Integrated Manager for Lustre software embedded in EXAScaler. Click a desired topic.

Dashboard

Dashboard is the first view that you see on starting a new session in the GUI. You can also navigate to this view from any other view by clicking Dashboard at the top or on the breadcrumbs trail.

md_Graphics/dashboard.png

You will see a portable, “sandwich” view for all browser windows less than 1024 pixels in width.

md_Graphics/dashboard-sandwich.png

The Dashboard view displays four charts with usage and performance data aggregated at several levels for monitored or managed filesystems:

Note that monitored filesystems do not support failover of their targets through the GUI.

Top of page

Login

If you are a filesystem administrator, log in to see a list of resources and the Management menu, as well as logs for the filesystem cluster. When you log in, besides seeing the default Dashboard view, you can select to monitor and manage individual filesystems and servers.

md_Graphics/dashboard-login.png

Top of page

Management menu

The Management menu appears at the top of the GUI page after login. From there, you can click a resource to jump to its details and do the various management actions on the resource. The list of resources on the left, which also appears after login, mirrors most of the Management menu and makes it easy to drill to a particular resource.

md_Graphics/mgmt-menu.png

Top of page

Breadcrumb navigation lets you see where in the hierarchy of the GUI you currently are.

md_Graphics/breadcrumbs.png

Breadcrumbs display a path outlining the steps taken to arrive at the current page, up from your starting point.

You can click a breadcrumb to jump directly to the corresponding previous step. For consecutive navigation, use the Back and Forward buttons of the browser.

If you create a cycle by clicking an item already seen and listed in breadcrumbs, the breadcrumbs automatically slice up to the current location, preventing an unnecessary build-up of items. If you drill down to an item and then refresh the page, the item becomes the only one in the breadcrumb list. Breadcrumb navigation is also reset if you click the Dashboard link at the top of the page.

Top of page

Servers

To see filesystem servers, log in to the GUI, and then on the left click Servers or click Management > Servers on the top.

md_Graphics/servers.png

You can identify the listed filesystem servers by:

Top of page

Server actions

There are several actions that you can perform on servers and related LNet in Servers view.

For any listed server, you can do:

You can additionally control the Lustre networking service:

There is also a linked host name for each server. Click it, and you will jump to a server-specific view with the same server and LNet actions. Alternatively, expand the list of servers on the left and click a server. Remember that you must be logged in to do this.

md_Graphics/server-details.png

Top of page

Server volumes

To see all volumes, for all servers, log in to the management interface, and then click Management > Servers on the top:

md_Graphics/server-volumes.png

Alternatively, expand the list of volumes for a server on the left:

md_Graphics/server-volumes-alt.png

If there are a lot of volumes, the list is paginated.

Volumes, also called LUNs or block devices, are the underlying units of storage used to create filesystems. Each filesystem target corresponds to a single volume.

Currently, you can only see the volumes that are in use as filesystem targets. If a volume’s server, also called host, is configured for HA, you can also see the volume’s peer servers. You can see that a volume may be accessible on one or more servers via different device paths, such as /dev/disk/by-id/scsi-360001ff0e06680000000002a880d000b, and that a volume may also be accessible via multiple device paths on the same server. In a future version of the management interface, you will be able to set or change the volume’s primary and peer servers.

Top of page

Server dashboard

Each server has a dashboard with four charts, which you can see if you log in to the management interface, expand the list of servers on the left, and click a chart icon next to a server’s name.

md_Graphics/server-dashboard.png

The server dashboard displays the following charts:

Server’s read/write bandwidth

The Read/Write Bandwidth chart on the server’s dashboard shows the read and write activity over time for the filesystem(s) that the server hosts.

md_Graphics/server-bandwidth-chart.png

Server’s CPU usage

The CPU Usage chart on the server’s dashboard shows the percentage of CPU activity over time, measured individually for user processes, system processes, and processes in an IO Wait state.

md_Graphics/server-cpu-chart.png

Server’s memory usage

The Memory Usage chart on the server’s dashboard shows total vs used RAM and total vs used swap space.

md_Graphics/server-memory-chart.png

Server’s LNet usage

The LNet Usage chart on the server’s dashboard shows packets received vs sent over each Lustre Network Identifier on the server, such as 172.18.0.1@o2ib.

md_Graphics/server-lnet-chart.png

Top of page

Filesystems

You can list filesystems in the cluster if you log in to the management interface, and then click Filesystems on the left or instead click Management > Filesystems on the top.

md_Graphics/filesystems.png

Each filesystem listed in Filesystems view is identified by:

Top of page

Filesystem actions

There are several actions that you can perform on filesystems in Filesystems view:

Top of page

Filesystem details

To see filesystem details, click the name of a filesystem in Filesystems view or, alternatively, expand the list of filesystems on the left and click the name of a filesystem. Remember that you must be logged in to do this.

md_Graphics/filesystem-details.png

You can scroll up and down the details, and additionally expand the list of filesystem resources on the left to use it for navigation.

md_Graphics/filesystem-details-nav.png

On the details view, you see filesystem details as well as details for filesystem resources, such as MGS, MDTs, OSTs, and OST pools.

The following details identify the filesystem in addition to the name:

The following details identify MGTs, MDTs, and OSTs in addition to the name:

Here is how the details may look like for a metadata target:

md_Graphics/target-details.png

If you drill down the list of filesystem resources to OST pools, you will see the OSTs of a pool, and can click any target to display the details as above.

md_Graphics/pool-details.png

You can perform actions on each target (MGT, MDT, or OST) that you view:

Top of page

Filesystem dashboard

Each filesystem has a dashboard with the charts:

To see the dashboard, log in to the GUI, expand the list of filesystems on the left, and then click the chart icon for a filesystem:

md_Graphics/filesystem-dashboard.png

Top of page

MGT dashboard

Each filesystem’s management target has a dashboard with the charts:

To see the dashboard, log in to the GUI, navigate to a filesystem listed on the left, expand the list of MGTs, and click the chart icon for a target:

md_Graphics/mgt-dashboard.png

Top of page

MDT dashboard

Each filesystem’s metadata target has a dashboard with the charts:

To see the dashboard, log in to the GUI, navigate to a filesystem listed on the left, expand the list of MDTs, and click the chart icon for a target:

md_Graphics/mdt-dashboard.png

Top of page

OST dashboard

Each filesystem’s object storage target has a dashboard with the charts:

To see the dashboard, log in to the GUI, navigate to a filesystem listed on the left, expand the list of OSTs, and click the chart icon for a target:

md_Graphics/ost-dashboard.png

Top of page

Users

If you are a superuser, you can view, create, and manage users by clicking Management > Users at the top of the page.

md_Graphics/config_users.png

The following types of users exist:

Superusers also have these options by clicking a user’s account:

Top of page

Logs

To view logs, log in to the management interface, and then click Logs on the right. The view will expand and display a paginated list of logged events.

md_Graphics/logs.png

You can identify each logged event by time, service, message, and host.

As an example of logged events, consider what happens if a failover takes place:

An event is logged for each case above, and you can see it displayed in Logs.

Top of page

Activities

Activity messages provide information about the functioning and health of a managed filesystem. You do not need to be logged in to view activity messages.

To view activity messages, click Activity on the right of the page. The view will expand and display a paginated list of all filesystem activities, listing most recent messages first.

md_Graphics/activities.png

The counter next to the activity icon reflects the number of active issues with the cluster. The color of the icon will change between green, yellow, and red according to the severity of the highest active issue.

There are five types of activity messages:

Before you see an unresolved filesystem error or warning in the Activity view, this error or warning pops up as a browser-level notification.

Top of page