zf7372 APs sending LLDP request for 130W of PoE Power to Switch

[james_maslen]

I have four zf7372 APs connected via CAT6 to a ProCurve 5406 switch. The switch has 275W of PoE power available, and in total only has five PoE devices. Switch ports are 802.3af PoE only, no 802.3at PoE+.

When the 7372 APs first power up the switch logs show "PD Denied power due to insufficient power allocation" for each AP. A few moments later, the AP requests PoE again but this time it works, no switch error is in the logs, and the AP powers on as expected. This pattern will repeat randomly throughout the day, which is resulting in APs going offline frequently.

After troubleshooting with the switch vendor we found the following by debugging LLDP:

0001:19:34:31.60 LLDP mlldpCtrl: lldp refresh pkt sent out port : A7
0001:19:34:31.60 LLDP mlldpCtrl: lldp refresh pkt sent out port : A8
0001:19:34:33.60 LLDP mlldpCtrl: lldp refresh pkt sent out port : A2
0001:19:34:38.14 LLDP mlldpCtrl: device status changed : 11
0001:19:34:38.14 LLDP mlldpCtrl: 8023at: PoeMgr PowerUpdate  to LLDP (mtype = LOC_SYS_CHG) on port A8
0001:19:34:38.14 LLDP mlldpCtrl: 8023at: PowerUpdate values PseAloc=24,PdReq=130 (Type=2, Source=1, Priority=3) on port A8
0001:19:34:38.14 LLDP mlldpCtrl: device status changed : 11

The area of concern is "PdReq=130" which seems to indicate that the Ruckus APs are initially requesting 130W of PoE power for each AP. A second set of LLDP refresh packets sent to the switch a few moments later contains a more reasonable value, which the switch then delivers and powers the AP without a problem until the pattern repeats.

I have exhausted all troubleshooting options with the switch vendor and would like to see what, if anything, can be done from the AP-side of troubleshooting or isolating this problem.

APs are controlled by an on-prem ZD1200 running software version 10.1.2.0.251. Switch is on latest available code as well (K.16.02.0028).