Cisco 3850 16.12.3a POE issues

[AdamF1]

Good morning,

Is anyone else running 16.12.3a IOS on 3850 switches?

Here is an issue we are facing but I can not find any documentation of a bug in this code and its still recommended as the code to go to...

Issue: POE Stops functioning on random ports but works on others. POE will not work for Avaya phones, cameras, cisco phones, or Cisco AP's ( 3602,3702,3802).

Work around: Reboot switch, downgrade, or find a port that will provide POE.

We began upgrading and testing on several stacks for a month or 2 with no issues prior to deploying to approximately 30 stacks of 3850's. After we did a mass deployment we began to see POE issues on switches that seem to be triggered when removing or adding a POE device. Once the condition has been triggered it will not go away until rebooted or downgraded. Logs will state " Controller port error, interface x/x/x, power given, but machine power good wait timer timed out.

I have found similar issues or bugs in older codes, have we regressed?

[Leo Laohoo]

Go straight to 16.12.7.  

If it is possible to cold-reboot (pull the power cable) the entire stack, make it happen.

[StevenCAnderson]

We were having this issue on some 3850's with 16.12.3a and 16.12.8. After trying a few things recommended here nothing worked. Although none of the ports were in an errdisable state, I ended up turning on errdisable detect cause inline-power, errdisable recovery cause inline-power, and errdisable recovery interval 30. After 30 seconds the ports all came up and pulled IP's and we haven't had any issues since. I put the commands on all the devices having the issue and I haven't had the issue since. Not sure what is really going on but this fix has worked for now. We will see how long it lasts. 

[oscar.garcia88]

Thanks for your comments Steven,

So, tyou are saying that he solution is to enable inline-power command on the ports that do not give PoE, and after 30 seconds the affected ports come up?

[StevenCAnderson]

The solution that worked for me was the error disable recovery function. I put the following commands on the switches having the issue:
errdisable detect cause inline-power      <-- This was on by default but if it wasn't then this turned it on.
errdisable recovery cause inline-power   <-- This allows the switch to recover the port if an inline-power issue triggers the errdisable state. None of the ports showed up as errdisabled, the switch didn't acknowledge them as being disabled.
errdisable recovery interval 30    <-- This sets the interval to 30 seconds, normally I do 600 so that recovery occurs at 10 minutes but to test it I set it to 30 seconds. 

I'm not guaranteeing that this fixes the issue for all the different POE bugs related to the 16.12.X track, but this so far has remained a satisfactory fix for a few months. There was also a different POE issue that was fixed by forcing the power statically. The device wouldn't show up via CDP even though it was plugged in, by putting power inline static max 15400 I was able to get the phones to show up as CDP neighbors.