Upgrading node firmware

Whenever a firmware upgrades available for nodes, run firmware upgrades on them.

Before you begin

Upgrade the IBM Spectrum Fusion HCI management software.
Post the upgrade of IBM Spectrum Fusion HCI management software, check whether firmware upgrade exists by using any of the following methods:
- Open the Red Hat® OpenShift® Container Platform web console. In the Overview page, go to Activity > Recent events section, see whether you can see the following message:
```
IBM Spectrum Fusion firmware upgrade available for nodes. 
```
- Open the IBM Spectrum Fusion HCI user interface displays notifications in the following locations:
  - The message is displayed in the Quick start page of IBM Spectrum Fusion user interface:
```
Firmware upgrade available
An upgrade (version <version_number>) is available for the nodes
```
  - In the bell icon, you can see the following messages for node and switch:
```
Compute firmware upgrade available
```
  - In the Infrastructure > Nodes, you can see the following message:
```
Version <version_number> firmware upgrade is available
```
If a firmware upgrade exists, the admin must perform the following checks before updating the firmware:
- Ensure that the node is part of the OpenShift Cluster.
- Ensure that none of the compute nodes in the cluster are in maintenance mode before triggering a firmware update.
- If previous attempt of firmware update failed, before you reattempt it, delete the previous computemaintenance CR.
  1. Run the following command to list all computemaintenance CR:
```
oc get cmt
```
    For example:
```
NAME                   AGE
firmware-compute-0            37h
```
  2. Run the following command to delete the failed firmware upgrade:
```
oc delete cmt <firmware-compute/control-X>
```
    For example:
```
oc delete cmt firmware-compute-0
```
- Ensure that you put a node into maintenance mode and that the application must be evacuated before you upgrade firmware, so that the application load is not impacted by this operation.

About this task

Do the node firmware upgrade in the specified sequence.

Procedure

Open the IBM Spectrum Fusion HCI user interface.
Go to Infrastructure > Nodes.
In the Nodes page, select the nodes that you want to upgrade.
As soon as you select the nodes, the Upgrade button is displayed.
Click Upgrade.
The Upgrade firmware on <n>nodes window gets displayed.
Click Start upgrade.
The following message appears to confirm the initiation of upgrade:
```
Upgrading firmware on <n> nodes
```
Though you select multiple nodes, they are upgraded one-by-one. Before the actual upgrade, the node is automatically moved to maintenance mode.

Note: If a firmware update fails for a node in each set of nodes, the remaining nodes are not updated until the node is fixed to a ready state.

The admin can check the status of the firmware upgrade under the Firmware tab on the Nodes page.

The following table lists the various node states that provide information about the firmware upgrade.

Firmware State	Description
<version>	Current firmware version
<version> - Upgrade Available	Indicates that an upgrade is available and displays the current firmware version.
<version> - Upgrade scheduled	Indicates that an upgrade is scheduled for the next firmware version.
<version> - Upgrade in progress	Indicates that an upgrade is in progress for the next firmware version.
<version> - Upgrade failed	Indicates that an upgrade is failed for the next firmware version.

If the firmware upgrade fails, then the following system notification is displayed.
```
Firmware upgrade failed
```
If the upgrade failed for any reason, collect the Compute logs and select the failed node to do the following steps:
1. If you want to reattempt the firmware upgrade, click Retry upgrade.
2. If you want to cancel the firmware upgrade, click Cancel upgrade.
Note:
The following are the tasks that the administrator must be aware of during the firmware upgrade.
- The admin can schedule firmware updates for a selected set of nodes.
- If a scheduled upgrade is selected, the firmware updates compute nodes one-by-one in each scheduled upgrade. The admin does not have any control over the order of node upgrades.
- The admin can schedule upgrades for more compute nodes during the upgrade process. Newly added compute nodes get upgraded according to their upgrade slot.
- If a firmware update fails for a node in each set of nodes, the remaining nodes are not updated until the node is fixed to a ready state. But the admin can retry the upgrade for the failed node.
- If a firmware update fails, a critical event is raised for that node. The critical event can be seen in the Recent Events or Events sections.
- The admin can cancel the upgrade for scheduled nodes only, but not for the node that is already in-progress.