← Back to blog

One kcore node from an evaluation ISO is a lab job. A rack of PowerEdge boxes is a different job. You want the machine list, an image you already checked, a boot target, and a way to turn the host on without walking the row with a keyboard.

kcore-provisioning keeps those on one controller. It talks to Dell iDRAC over IPMI and Redfish, and it netboots kcore with UEFI HTTP boot. A ThinkCentre has no BMC, so the same controller sends a Wake-on-LAN magic packet to the NIC. The dashboard is the operator view: images, machines, boot configs, and power. The screenshots are from the lab. Look at the layout.

What the dashboard shows

Overview is the fleet page. It counts OS images, how many of those are verified, and how many machines are provisioned, still installing, or failed. It refreshes every ten seconds from the controller API. The links go to Images, Machines, and Boot Configs. Recent activity lists the latest events when there are any.

Open Overview. That is the list of hosts that are ready.

Feature list

OS Images table with three verified kcoreos images
Images: verified kcoreos builds ready to assign as boot targets.

Why a fleet is easier this way

Doing this by hand falls apart after a few nodes. Someone images each disk, sets the boot order, and writes down which MAC got which build. The provisioner shortens that loop:

You pick the image, the machine is already in the inventory, and you power it on. The hardware is the hardware you already racked.

Machines table listing three Dell PowerEdge T140 servers
Machines: three Dell PowerEdge T140 hosts in DISCOVERED state with boot targets assigned.

A machine page in practice

Open one host. Power sits next to drive health. CPU, memory, BMC firmware, NICs, and disks are on the same screen. The boot target is there too, and you can download the config. You stay off the iDRAC page, the PXE UI, and the wiki note that used to hold the kernel line.

Machine details with power control and BMC information
Machine details: power control, BMC data, NICs, storage, and boot target on one page.

Where this sits in kcore

After a node is installed, VMs and clusters still go through kctl, YAML, and the Terraform provider. The provisioner is how the bare metal gets into that fleet. One host or a small lab still starts from the evaluation ISO. A rack starts from the provisioner.

For IPMI, Redfish, and UEFI HTTP boot, see features. For node join and approval after install, see the add-node and installation guides.

Conclusion

The machine list, the images, the boot target, and the power button used to live in four tools. They now live on one controller and one dashboard. A checked image, a discovered machine, a boot config, and either the BMC or a Wake-on-LAN packet is enough to bring the next server up.

Download the evaluation ISO See features Installation guide See pricing

Questions: team@tacconiconsulting.com.