Skip to content

Work with many devices at once

The Devices table acts on a group of devices in one step: onboard, undeploy, set a credential or a role, or replace the images they stage.

In the Console

Select devices

  1. Open Inventory.
  2. Click a row to select it, and click again to deselect it. Selected rows are highlighted and show the bulk actions.
  3. Read the selected count before you act on it.

Select page and Clear page selection act on the visible page, Select all N matching devices adds every matching device, and Cancel clears everything. Selections stay selected across pages and filter changes. The bar also counts selections outside the current page or filter. Use Cancel first when you want to start a different group.

Select all reads every page before adding devices. If a page cannot be read, nothing is added. Changing filters, changing the selection or leaving Inventory cancels that read.

Search and filter

Search by device, IP or model. Filters opens the structured filters, and Reset clears them. A filter applies on the server, so it covers every matching device, not only the page in front of you. Device series lists IE Switches, IR Routers, Catalyst Routers, Catalyst Switches, NCS and Cisco 8000 Series, with legacy and unknown categories retained. Hover a series cell to see the stored model. For the role filter and Set role, see Control which devices share with each other.

The table refreshes every ten seconds while Inventory is visible. Your filters, selection and keyboard focus stay in place. If the device list cannot be read, the selection remains, but its match to the current filter is unconfirmed.

Credential and sharing-policy controls are disabled when their current data is unavailable. If an edit reports an unknown outcome, check the refreshed row before trying again.

Run a toolbar action

Action Effect
Onboard Queues an agent deployment for each selected device.
Undeploy Removes the agent using its deployment record.
Adopt Records a reviewed existing deployment.
Delete Retires inventory and server-side device state.
Set credential or role Updates the selected devices. A role change previews first, so read the impact counts before you commit it.
Assign images Replaces each selected device's image set with the checked images.

Warning

Delete leaves the agent on the device. Undeploy the devices first, as described in Undeploy, retire and clean up devices.

The toolbar acts on every selected row. Use Adopt when an agent is already on the device and no record covers it. Schedule… targets the current filter, not the rows you selected, see Schedule maintenance windows.

Assign a set of images

  1. Select the devices and click Assign images. The picker opens pre-checked with the images assigned to every selected device.
  2. Check the images the whole selection should stage, up to ten.
  3. Apply. Different assignments are flagged. Removing existing assignments asks for confirmation.

The dialog keeps the devices you reviewed when you opened it. For conflicts or an unknown result, follow Assign images and check staging status.

Warning

Applying replaces each device's whole set and drops anything left unchecked, so it can also remove assignments. An empty set unassigns every image on the selected devices. See Assign images and check staging status.

With the API

For the same work from a script, see Automate with the API. The device routes and their schemas are in Console API.

What you see

The deployment details drawer

Each row has a ⓘ Deployment details control. It opens a drawer that lists every image assigned to that device. Esc or ✕ closes it.

If a re-imaged or replaced device reports a changed SSH host key, verify its identity before choosing Forget host key and confirming. The next connection trusts and saves the key it presents. Success confirms that device's saved entry is absent, including hashed entries; other devices' keys stay unchanged. If the request outcome is unknown, check the saved key before trying again.

Saved deployment logs appear below the record. Missing dates and results show as unknown. Previous device marks a log dated before this device was added.

State What it means
ready Staged and verified.
error The agent reported a failure.
staging The agent is working on the image.
pending Staging has not been reported.

For one image, the drawer shows the agent's detailed state and error. For several, the shared error appears in Last reported error. The Status column shows N of M image(s) failed.

The deployment record

Below the images table, the drawer shows the record id and its state: active, removed, superseded, needs-reconcile or abandoned. An abandoned record no longer describes a device IRIS manages, and a fresh onboard still works. A device with no record names adopt or a fresh onboard as the fix.

It also shows the preflight result and what the onboard applied: the management type, the owned management VLAN or VPG (VirtualPortGroup), the SVI (switched virtual interface) and app addressing, the NAT interface, the swarm port, the recorded model, the Agent install choice and the device identity. Management type is how the agent reaches the network: on its own address, on your management VLAN, or through the router, see Choose a management type. Agent install is where the agent runs: on Catalyst 9000 series switches the agent runs in Guest Shell, and the IOx app is the alternative on switches with app-hosting storage.

The drawer ends with that device's deployment logs, viewable in place. Logs are keyed on the device id and survive a delete. A run from a box registered earlier under this name is labeled previous device.

When part of a bulk action fails

A bulk action reports each success and refusal instead of failing the whole batch. The status line counts the successes and names the devices that refused. Delete also names partial or unknown results and keeps those devices selected; see Undeploy, retire and clean up devices. HTTP 207 means partial cleanup. HTTP 429 means the server rejected the request on admission, and the response carries Retry-After. Those limits are in Console API. For a refusal you cannot place, see Troubleshoot: symptoms and first steps.