The Cluster section provides access to the details of a cluster's Hardware (physical or virtual machines, or chassis), Subclusters, and Feeds (search and replication).

Streams versus Objects

Streams are not objects. One object can involve many streams. Objects are the files uploaded to and downloaded from Swarm, and they are indexed for searching and listing. Streams are all underlying data components Swarm manages internally, such as erasure-coded segments, replicas, historical versions, and chunks of multipart uploads. The Swarm UI reports on streams when it is surfacing Swarm's internal view of hardware usage.

Dashboard

The primary initial view of the cluster is offered through the Dashboard. The Dashboard presents real-time visual monitoring, alerting, and history across the cluster's usage and activity. The host name and last refresh time for the page are shown at the top. The page auto-refreshes every 60 seconds.

Empty Dashboard

The top of the page shows a Swarm watermark background if the cluster does not have an active license. Attempt to log out and log in again if the license is active but the data is still missing.

The Dashboard provides quick reference to these key indicators for the cluster:

Health

Cluster

Overall cluster health as a red, yellow, green status based on the status of the cluster's chassis and disks. Drill down to the Health Report for additional details.

Tip

Hover over sections to read the values in popup text if unable to distinguish among those colors.

Note

Per-object feed errors (for replication or search indexing) do not generate critical log messages and so do not change the node state to 'error'. Check the Feeds section of the Dashboard for blocked feeds.

Usage

Disk Space

% of Disk space usage in the cluster, including free space, used space and trapped space. The Disk space pie chart turns orange if the amount of used capacity exceeds 80% and then red if the capacity exceeds 90%. Mouse over each segment in the pie chart for more information.

Stream Index

% of Stream index usage in the cluster, including how much is used by the overlay index if any. The Stream index pie chart turns orange if the amount of used RAM exceeds 80%, and then red if the capacity exceeds 90%. Mouse over each segment in the pie chart for more information.

Elasticsearch

Statistics

The Elasticsearch panel provides a snapshot of the ES cluster's status and essential statistics. Click the icon for the full report for details.

Feeds

One pane per feed

Status for each feed, both search and replication, including current feed status, a trended graph of feed event processing over time, and a count by stage (Pending, Processing, Retrying) of any queued events and deletes. An animated spinning icon displays to indicate the feed is currently processing if queue for events or deletes is greater than zero.

Data Delays

It may take over 24 hours before the processing trend is visible in the graph.

Spinners appear when the UI is waiting on feeds still processing data, but they also appear when the feed itself is busy.

Network Traffic

Request

The count of each SCSP method type in incoming client requests over a rolling 30 day window.

Responses

The count of HTTP response codes returned to clients by the storage cluster over a rolling 30 day window.

The collapsible global menu pane provides navigation to more detailed information and reports. Use the icons on each section of the dashboard to drill down in to details for that section. Clicking the medical bag icon in the Dashboard's Health section opens the same page as selecting Reports > Health from the menu:

Hardware

The Hardware report includes a summary view of each server chassis in the storage cluster and the current state, including the number of chassis disks online, the used capacity, stream count, up-time, Swarm storage software version, and the subcluster to which the chassis belongs, if any.

Click on any row to drill down in to the Chassis Details page for that particular chassis (physical or virtual machine):

Restarting or Shutting Down the Cluster

The settings gear icon at the top of the page allows restarting or shutting down the entire storage cluster, as well as the ability to clear logs.

Admin Only

These cluster-wide actions require authentication as an administrator.

Perform a Rolling Restart of the cluster with full control over the restart queue, to reorder and cancel individual chassis restarts if Platform Server is installed. (v2.0)

Suspending or Enabling Disk Recovery

Administrators can suspend an in process disk recovery using the Suspend Recovery option under the settings gear icon.

Enable a previously suspended recovery using either the Enable Disk Recovery button in the banner message or the Enable Recovery under the settings gear icon.

Subclusters

The Subclusters page rolls up the information on all subclusters present in a cluster, including chassis and disk counts and number of streams.

Navigate to Hardware > Chassis Details and click the Settings tab to dynamically change a subcluster assignment. Subclusters can be configured from the CSN (cluster.cfg) or in the node's configuration file (node.cfg), but these require a cluster reboot to take effect.

Swarm Documentation

Viewing and Managing the Cluster