Skip to content

Update Host Firmware (Hardware Lifecycle Manager)

The console reads and updates host firmware (BIOS, BMC, NICs, RAID/HBA controllers, disks, power supplies and so on) through the host's BMC, using the standard Redfish Update Service. It works with Dell iDRAC, HPE iLO, Lenovo XClarity and other Redfish-compliant BMCs, and needs no separate vendor management tool.

Connect the host's BMC

Right-click the host → Hardware Health. If no BMC is configured, click Configure BMC / iDRAC, enter the BMC / iDRAC IP Address, Username and Password, then Save & Connect. These are the same credentials HA uses to fence the host.

See what a host runs

Hardware Health → Firmware lists every component with its version and whether it can be updated.

To compare a whole cluster, open the cluster's Summary → Firmware Compliance (HLM). It shows which components differ between hosts, and which hosts run which version. All nodes agree means the cluster is consistent.

Update one host

  1. Put the host in Maintenance (right-click → Enter Maintenance), so its VMs move off first. The console refuses an update on an Active host, with one exception: an update to the BMC itself is allowed while the host is Active, because it only restarts the BMC, not the host.
  2. Get the firmware image from the hardware vendor, check it against the vendor's published checksum, and put it on a web server the BMC's management network can reach.
  3. Hardware Health → Firmware → Submit Firmware Update: pick the Component, enter the Image URL (the BMC downloads the file itself) and, optionally, the Target Version for your records.
  4. Submit Update and confirm. This is a real update, not a test.

Status and Progress in Update History follow the BMC's own job until it finishes. The page doesn't need to stay open. When it's done, check the new version in the list, then Exit Maintenance.

Reboots

Many components, such as BIOS and NIC firmware, apply only when the host restarts. The BMC decides when, depending on the package. Keep the host in Maintenance until the update history says it has finished and the new version shows.

Update a whole cluster (firmware campaign)

A campaign brings a cluster's hosts to the same firmware, one host at a time, without further clicks.

  1. Cluster Summary → Firmware Compliance (HLM).
  2. Tick each component to bring into line, pick its target version (the list shows how many hosts run each), and enter the Image URL for target version.
  3. Start Campaign and confirm.

For each host in turn, the console:

  1. moves its VMs to other hosts (Maintenance),
  2. submits the updates and waits for them,
  3. checks the new versions,
  4. returns the host to service,

and only then moves to the next host. The card shows each host's progress. Pause, Resume and Cancel act on the whole campaign, and Skip node passes over one host. A campaign keeps going through a console restart.

The campaign stops rather than guessing:

  • If the rest of the cluster doesn't have room for a host's VMs, the campaign fails as a whole. Free up capacity and start a new one.
  • A host running a VM on local disk, which can't be live-migrated, halts the campaign for you to deal with that VM. Move or shut it down, or Skip node.
  • If evacuating a host, the update itself, or bringing the host back fails, the campaign halts on that host with the reason. Fix it, or Skip node to carry on with the rest.

Try it on a test cluster first

Firmware packages and BMCs differ in how they apply updates and when they reboot. Run your first campaign with a given vendor's packages on a non-production cluster, or update one host by hand first and watch how long it takes.

See also