PoE Budget and Uplink Errors on an EX Switch Stack
A logistics company has 45 Juniper EX4300-48P switches managed by Mist. One warehouse IDF is a four-member stack serving 24 APs (802.3at, 25.5 W each) and 40 cameras (802.3af, 12.9 W each), plus 60 access ports. The stack has a 740 W PoE budget per member. Member 3 currently powers 24 APs and 16 cameras; members 1, 2, and 4 each power 8 cameras. Since last week, APs on member 3 reboot randomly, Mist Wired Assurance shows PoE budget alarms, and the stack uplink reports CRC errors. The team replaced one AP patch cable but the reboots continue. Mist shows the member 3 uplink at 1 Gbps with 0.8% CRC errors and intermittent port flaps. The company wants to avoid adding another switch if possible.
1. Calculate whether member 3 is over its PoE budget and identify the most likely cause of the AP reboots.
Read the worked answer
Member 3 powers 24 APs at 25.5 W, which is 24 × 25.5 = 612 W, plus 16 cameras at 12.9 W, which is 16 × 12.9 = 206.4 W. Total PoE load is 818.4 W, or about 78.4 W above the 740 W budget. With any realistic reserve for cable loss and surge, the member is even further over budget. That matches the Mist PoE budget alarm. The AP reboots are most likely PoE port power shedding or overload behavior on member 3, not a bad AP patch cable, because the load exceeds the member budget and the reboots continue after one cable was replaced. The uplink CRC errors are a separate physical-layer symptom and should not be attributed to PoE without checking the cable, optics, and port.
2. What Mist Wired Assurance data and switch configuration changes would you use to remediate?
Read the worked answer
Use Mist Wired Assurance to view per-port PoE consumption, PoE budget alarms, port flaps, and the switch health timeline for member 3. Confirm which ports are delivering 802.3at versus 802.3af and whether any unused ports still have PoE enabled. Rebalance load by moving approximately 8 APs from member 3 to members 1, 2, or 4, which each currently carry only 8 cameras, or add a supported power supply if the platform allows more PoE capacity. Disable PoE on unused access ports and apply port profiles through a Mist switch template so the change is repeatable. Give AP ports higher PoE priority than cameras if the platform supports it, so cameras are shed before APs during a transient budget event. Validate after the change that member 3 is below 740 W with headroom and that the PoE alarms clear.
3. How do you separate the uplink CRC errors from the PoE issue, and what would you check?
Read the worked answer
Treat CRC errors as a Layer 1 problem independent of PoE. Check the uplink cable, connectors, patch panel, and optics; run a cable test if supported; and inspect the interface error counters before and after replacing the cable or optic. Confirm speed and duplex are as expected and that the port is not flapping because of a duplex mismatch or a faulty SFP. Correlate the flap timestamps with the AP reboot timestamps: if APs reboot only on member 3 while the uplink flaps independently, the two are likely separate. If the uplink flaps cause stack control-plane instability, fix the uplink first. After remediation, monitor CRC counters and PoE budget alarms together to confirm both issues are resolved.