Description

When Redfish firmware/management/RAID operations fail, some BMC implementations (notably HPE iLO) return only a MessageId field (e.g., "iLO.2.34.DailyUpdateLimitExceeded") without providing a human-readable message field in the task response. This is a valid Redfish response format per the specification.

The current implementation extracts error messages using:

  
  messages = [m.message for m in sushy_task.messages]

When m.message is None/empty, this results in empty strings in the messages
list, causing:

• Empty error messages displayed to users
• Misleading "OK" status indicators in Bare Metal Host (BMH) resources
• No actionable information for troubleshooting failures

Users cannot diagnose firmware update, management, or RAID configuration failures when their BMC only provides MessageIds. This affects operations on HPE iLO and potentially other BMC vendors that follow this valid Redfish pattern.

Reference
https://bugs.launchpad.net/ironic/+bug/2136279