Low Power on Cisco 3802i Access Point

[Jerome N]

Hi,

I am currently trying to troubleshoot why a 3802i is staying in a low powered mode even with a 802.3at PoE injector attached. It is a deployment of around 12 APs with 6 of them remaining in that state.

Logs on the APs show CDP PSE negotiation timeouts. They also are showing "AC Adaptor (false)...802.3af (true).

I'm not sure how this is happening as the injectors are AT compliant. They are connected to a controller so all obtained the same global config. Same injectors used.

4 of the 6 worked initially but reverted to low power after dropping offline and having to be rebooted to regain management visbility. The other 2 have been in thay state from the beginning.

Other deployments have had this issue but started working once we rebooted by disconnecting the PoE.

I just feel the working ones are running on borrowed time and will fail eventually.

Hoping someone can shed some light on this. This involves a major customer and I would be extremely grateful for any assistance.

Jerome 

[Leo Laohoo]

1. What is the exact model of the Injector?
2. What firmware is the controller running on?

[Jerome N]

1. Microsemi 9001GR
2. 8.5.103.0 on a 2504 WLC

[Leo Laohoo]

Remote into the AP and post the complete output to the command "sh cdp inline".

[Jerome N]

AP13#sh cdp inline
Power_Requested(mW) Power_Available(mW) Power_request-ID Power_management-ID
Not-negotiated Not-negotiated


Same output on all the APs.

[Leo Laohoo]

Oh geez. I forgot, this is Microsemi. Me bad.
Can you please post the complete output to the command "sh lldp inline".

[Jerome N]

No worries.

AP13#sh lldp inline
Power_Requested(mW) Power_Available(mW) Port_ID
Negotiation-failed Negotiation-failed

A working AP:

AP03#sh lldp inline
Power_Requested(mW) Power_Available(mW) Port_ID
Not-negotiated Not-negotiated

[Leo Laohoo]

Oh WTF.
Ok, next trick. Please upgrade the firmware of the controller. Try using 8.5.110.0.

[Jerome N]

Will do.

Will have to come back to you in a few days if that's alright. Not permitted any changes over the weekend as there is no support.

[Leo Laohoo]

No problem. Let us know the situation after the upgrade.
Do remember that if you're WLC is a 250/5508/WiSM2, there is a Supplementary AP Bundle necessary if you've got some legacy APs.

[Jerome N]

I'm not sure. Only remember the model being 2504.
We recently upgraded them all to 3802i so there are no longer any legacy APs.

[Jerome N]

Hi,

Thought i'd update this for any one that reads this and has a similar issue. I found in another thread a way of rebooting the access points to get them out of low power mode:

1. Disconnect the data uplink from the injector to the switch.
2. Reboot the injector with the AP still connected.
3. Wait for the indicator on the injector to turn green and then reconnect the uplink.

Note: If the indicator does not turn green, this process will not work.

This has allowed the APs to obtain full power. My understanding of this is that leaving the cable disconnected when the AP boots up allows for it to establish that the injector is the power source. With the uplink connected, I think the AP knows there is a switch there and is awaiting CDP/LLDP negotiation of power hence the negotiation timeouts.

We are currently ordering the exact model of PoE injector that Cisco defines should power 3800s. It's annoying that it was never stated that other 802.3AT injectors may not work but I guess these things happen. I'm hoping the new injectors should eliminate the issue altogether.

Jerome

[Leo Laohoo]

---

@Jerome N wrote:
We are currently ordering the exact model of PoE injector that Cisco defines should power 3800s. 

---

Order 1 first to see if it works or not.  

[lukewarms]

Hi Leo,

Is upgrading to 8.5.110.0 or beyond (proposing 8.5.140.0) is supposed to fix this? We have a 5508 running 8.2.166.0 with 3802i's and experiencing the same issue.

Thanks

Luke

[Leo Laohoo]

Give the NEW firmware a try.