Symptoms:
Run the following commands to confirm:
Check PoE port status.
get switch poe inline
Output shows all ports Disabled:
Unit Power Budget: 800.00W
Unit Guard Band: 6.00W
Unit Power Consumption: 0.00W
Interface Status State Max-Power(W) Power-consumption(W) Priority Class Error
------------------------------------------------------------------------------------------------------------
port1 Disabled Disabled 0.00 0.00 Low 0
port2 Disabled Disabled 0.00 0.00 Low 0
...
port48 Disabled Disabled 0.00 0.00 Low 0
Check per-port PoE diagnostic.
diagnose switch poe status port1
Output shows:
Port(1) Power:0.00W, Power-Status: Disabled
Power-Up Mode: Normal Mode
Remote Power Device Type: PD None
Power Class: 0
Defined Max Power: 0.0W, Priority: Low.
Voltage: 0.00V
Current: 0mA
Error while reading Temperature
The 'Error while reading Temperature' message on every port is a key indicator — the PSE controller cannot be read at all.
Check hardware status.
get hardware status
Critical line:
PSE Controller: BCM59XXX (0xff)
The (0xff) return value means the PSE controller chip is not responding to bus reads. When the controller returns 0xFF on all register reads, the poed daemon cannot enable power on any port.
Check PoE event logs.
execute log filter category poe
execute log display
Output is expected to show hundreds of poe-debug warning entries from poed across all ports, each containing 0xFF in the hex payload, confirming the bus communication failure.
Root cause:
The Broadcom BCM59XXX PSE (Power Sourcing Equipment) controller is the IC responsible for managing PoE power delivery on all ports. When it returns 0xFF on all register reads, it means:
The PSE controller is not being detected on the I2C/SPI bus. The poed daemon cannot query or set power state on any port. All ports default to Disabled as a safety measure. Temperature sensors (on the same controller) are also unreadable. There should be no config switch poe-global block (or it should be at default values). Individual ports should not have set poe-status disable unless intentionally configured (e.g., uplink/MCLAG ports). All other ports should have no poe-status override - they default to enabled.
If the config is clean but all ports show Disabled, the issue is hardware.
Troubleshooting steps:
Step 1. Reboot the switch:
execute reboot
A cold reboot may temporarily restore PSE bus communication if the issue is a transient I2C lockup. After the switch comes back:
Run get hardware status - if PSE Controller: BCM59XXX (0xff) persists, the hardware is confirmed faulty. Run get switch poe inline - if all ports still show Disabled, proceed to RMA.
Step 2. Verify power supplies.
diagnose sys psu status
Both PSUs should show OK. If either PSU is faulty, PoE may be disabled to protect the power budget. Replace the faulty PSU first.
Step 3. Check the peer switch (MCLAG setups).
If this switch is part of an MCLAG pair, run get switch poe inline on the peer switch:
If the peer also shows all Disabled with 0xff, investigate environmental factors (power feed, rack grounding). If the peer works fine, this is a single-unit hardware failure.
Step 4. File an RMA.
If the PSE controller still returns 0xff after reboot and both PSUs are OK, the switch requires hardware replacement (RMA). Include the following in the RMA case:
diagnose debug report output. get hardware status showing PSE Controller: BCM59XXX (0xff). get switch poe inline showing all ports Disabled. Confirmation that both PSUs are OK.
Related commands:
Command | Purpose |
|---|
get switch poe inline | View PoE status for all ports. | diagnose switch poe status port1 | Detailed PoE status for a single port. | get hardware status | Check PSE controller detection. | diagnose sys psu status | Verify power supply health. | execute log filter category poe | Filter logs for PoE events. | execute log display | Display filtered PoE logs. |
| 
