How do I check whether a dedicated server’s redundant power supplies are healthy?
Two installed power supplies do not automatically provide two independent power paths. Redundancy depends on supported PSU configuration, sufficient capacity and how each feed is connected. Management-controller telemetry can show PSU presence or faults, while the provider must confirm physical cabling and upstream feed independence.
Before you start
- Authorised management-controller access and provider confirmation of physical power connections.
Instructions
1. Read platform power health
Where the server supports local IPMI, inspect chassis and power-related sensor state. Names and redundancy indicators are vendor-specific.
sudo ipmitool -I open chassis status
sudo ipmitool -I open sdr type "Power Supply"These are reporting commands. A healthy reading does not independently verify which rack power circuit supplies each PSU.
2. Match telemetry to the physical plan
Ask the provider to confirm PSU models, redundancy mode, per-feed capacity and A/B cabling against the ordered service. Review any power-cap or high-load condition that could exceed single-PSU operation.
Check the result
- Both required PSUs are healthy, the intended redundancy mode is active and the provider’s verified feed map supports the required failure scenario.
Troubleshooting
- One PSU is absent or failed
- Assess whether the remaining supply can carry the workload and arrange model-specific replacement through the provider.
- Both PSUs use the same upstream circuit
- The server may tolerate one PSU fault but still lose power with that circuit. Document the actual protection scope.
- A failover test is proposed
- Coordinate a formal maintenance plan; do not pull a live feed as an informal test.
Official references
Use the documentation for your installed version when a flag, service name or platform feature differs.