Solved: 2960X - power inline faulty on interface

kimek4901 · ‎12-14-2016

Hello,

I have a switch 2960X with a power problem. It looks like two interfaces Gi 1/0/41 and Gi 1/0/42 (upper, lower) goes into power faulty state when Cisco 7911/ 7942 is connected. Log show TStart error on those ports even after 2xmode is applied on that port (I`ve already tried to reboot switch with no luck). I`ve checked it on two IOS:

c2960x-universalk9-mz.152-2.E4

c2960x-universalk9-mz.152-2.E5

error occured after 8-9 months after installation of 2960X switches. I have second switch with similar problem but there is no log and no PoE detection but it look the same, so two neighbour ports are not working with PoE.

Is is some 2960X bug? Anyone had the same issue and have a solution for that maybe?

Thanks in advance.

Marcin.

paul driver · ‎12-14-2016

Hello

Do the other POE ports work okay? - As I have had individual poe ports go on me and stop supplying any power, Nothing much I could do about it either, other than use POE injectors or swap the whole switch out which I wasn't wanting to do as the other ports worked fine?

res
Paul

Please rate and mark as an accepted solution if you have found any of the information provided useful.
This then could assist others on these forums to find a valuable answer and broadens the community’s global network.

Kind Regards
Paul

View solution in original post

Mark Malone · ‎12-15-2016

Yes agree with Paul I have ports all over our global network with descriptions set as POE dead , they go now and again on different platforms and blades , its whether you can live with it or RMA it and replace it, if its only one or 2 I usually just put up with it once there are more spare ports available

View solution in original post

paul driver · ‎12-14-2016

Hello

Do the other POE ports work okay? - As I have had individual poe ports go on me and stop supplying any power, Nothing much I could do about it either, other than use POE injectors or swap the whole switch out which I wasn't wanting to do as the other ports worked fine?

res
Paul

Please rate and mark as an accepted solution if you have found any of the information provided useful.
This then could assist others on these forums to find a valuable answer and broadens the community’s global network.

Kind Regards
Paul

kimek4901 · ‎12-15-2016

Hi Paul,

Yes, other ports are working fine, only two goes into faulty state of poe. I thought that it`s a hardware issue on those two switches and I`ll have to replace them but I wanted to try my luck and submit a question :)

Mark Malone · ‎12-15-2016

Yes agree with Paul I have ports all over our global network with descriptions set as POE dead , they go now and again on different platforms and blades , its whether you can live with it or RMA it and replace it, if its only one or 2 I usually just put up with it once there are more spare ports available

kimek4901 · ‎12-16-2016

Thank you guys for your time and sharing knowledge

Regards,

Marcin

sharlino · ‎02-17-2017

It seems I have same trouble as you and others.

Besides, switch does not show any errors in log and no ports are in faulty state in my case. In my opinion there is no guarantee that RMA will help, unfortunately.

c-ddeitch · ‎11-19-2018

I have the same problem as the others where POE stops working around 8 months none of my ports are working. I have upgraded the switch recently to 15.2(04)E06 with no luck. All ports show no poe.

"show power inline police <int-id>" shows Oper Status as "Off"

In our case issue was with MOSFET on Power Controller.

MCU I2C bus which hosts the power controllers which are responsible for POE. Each switch has either 6 (for 24 port SKU) or 12 (for 48 port SKU) power controllers (4 POE ports map to one Power controller).

For 2960X 48 port switch Power Controller 0-11

For 2960X 24 port switch Power Controller 0-5

Affected port was Gi2/0/23 on a 2960X 48 Port switch, Port 2/0/23 comes under power controller #6, however as numbering starts from 0 to 11 thus we used #5 in the command.

Enable "service internal" to run the frontend controller command.

2960-1#conf t

Enter configuration commands, one per line. End with CNTL/Z.

2960-1(config)#service internal

2960-1(config)#exi

2960-1#

2960-1#test frontend-controller read-poe 5

2960-1#

Jul 28 18:20:18.364: basic reg:

Jul 28 18:20:18.364: 0x08 0xF6 0x00 0x00 0x08 0x08 0x00 0x00

Jul 28 18:20:18.364: 0x00 0x00 0x00 0x00 0x00 0x24 0x0E 0x06 <---- 0x0E points to issue with MOSFET

Jul 28 18:20:18.364: 0x22 0x14 0xA4 0x0F 0xCC 0x00 0x00 0x80

Jul 28 18:20:18.364: 0x00 0x00 0x00 0x43

Jul 28 18:20:18.364:

Jul 28 18:20:18.406: next reg:

Jul 28 18:20:18.406: 0x80 0xE4 0xEF 0xFF 0x00 0x00 0x4C 0x00

Jul 28 18:20:18.406: 0x52 0x38 0x00 0x00 0x00 0x00 0x51 0x04

Jul 28 18:20:18.406: 0xED 0x36 0x00 0x00 0x00 0x00 0x00 0x00

Jul 28 18:20:18.406: 0x00 0x00 0x00 0x47 0x16 0xE2

Jul 28 18:20:18.406:

2960-1#

Has anyone else report this same issue? Did anyone find out for sure if this is hardware or software error?

Thanks

mario.jost · ‎07-08-2019

We have similar issues and try to find out how we can recognize the ports affected. I very much appreciate you posting the hidden stuff about this command:

test frontend-controller read-poe

But we get a very different output on our devices than you do. Is there a chart somewhere on what these Hex codes (0xF6 and so on) represent? I googled alot but wasnt successful in finding any documentation from Cisco for that matter. Only other command i found was this:

show controllers power inline module 1

But this does not show the defective PoE Ports. We have a switch in our lab that has port 1-8 (Power controller 0 and 1) defective and we want to test commands that could point these ports out on the CLI. We can then write a script that tests all ports in every country, rather than driving out each brach offices and testing the ports ourselfes.

Your answer is very much appreciated. Thanks and best regards.

Mario

ptst · ‎08-12-2019

Hello,

Me too; several ports on several switches, working with a Cisco engineer on how we could identify failing switches for sure.

I think that if you talk about Gi2/0/23 in a cluster, you may have to specify on which module you launch the test.

I see 2 ways to do that; the 1st isn't working for me actually, so I am using the 2nd one :

1) "session"

session 2
(user, password)
test frontend-controller read-poe 5

2) "remote command"

remote command 2 test frontend-controller read-poe 5

In our case, for failing ports Gi3/0/(27,28,30,31,32), here is the relevant output :

STACK_USER#remote command 3 test frontend-controller read-poe 6
Switch : 3 :
------------


STACK_USER#
Aug 12 17:17:16 CEST: basic reg: (STACK_USER-3)
Aug 12 17:17:16 CEST: 0x08 0xF6 0x00 0x00 0x0F 0x0F 0x00 0x00 (STACK_USER-3)
Aug 12 17:17:16 CEST: 0x00 0x00 0x00 0x00 0x03 0x07 0x06 0x06 (STACK_USER-3)
Aug 12 17:17:16 CEST: 0x00 0x20 0xAA 0x0F 0xFF 0x00 0x00 0x80 (STACK_USER-3)
Aug 12 17:17:16 CEST: 0x00 0x00 0x00 0x43 (STACK_USER-3)
Aug 12 17:17:16 CEST: (STACK_USER-3)
Aug 12 17:17:16 CEST: next reg: (STACK_USER-3)
Aug 12 17:17:16 CEST: 0x80 0xE4 0x00 0x00 0x00 0x00 0x52 0x00 (STACK_USER-3)
Aug 12 17:17:16 CEST: 0xC4 0x37 0x00 0x00 0x00 0x00 0x00 0x00 (STACK_USER-3)
Aug 12 17:17:16 CEST: 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 (STACK_USER-3)
Aug 12 17:17:16 CES
STACK_USER#remote command 3 test frontend-controller read-poe 7
Switch : 3 :
------------


STACK_USER#
Aug 12 17:17:19 CEST: basic reg: (STACK_USER-3)
Aug 12 17:17:19 CEST: 0x08 0xF6 0x00 0x00 0x0E 0x0E 0x00 0x00 (STACK_USER-3)
Aug 12 17:17:19 CEST: 0x00 0x00 0x00 0x00 0x0E 0x06 0x06 0x06 (STACK_USER-3)
Aug 12 17:17:19 CEST: 0x00 0x24 0xAA 0x0F 0xFF 0x00 0x00 0x80 (STACK_USER-3)
Aug 12 17:17:19 CEST: 0x00 0x00 0x00 0x43 (STACK_USER-3)
Aug 12 17:17:19 CEST: (STACK_USER-3)
Aug 12 17:17:19 CEST: next reg: (STACK_USER-3)
Aug 12 17:17:19 CEST: 0x80 0xE4 0x00 0x00 0x00 0x00 0x52 0x00 (STACK_USER-3)
Aug 12 17:17:19 CEST: 0xCA 0x37 0x00 0x00 0x00 0x00 0x00 0x00 (STACK_USER-3)
Aug 12 17:17:19 CEST: 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 (STACK_USER-3)
Aug 12 17:17:19 CES
STACK_USER#

mario.jost · ‎08-16-2019

Thanks for your post. I was more interested in how to interpret the results like:

Aug 12 17:17:16 CEST: 0x08 0xF6 0x00 0x00 0x0F 0x0F 0x00 0x00 (STACK_USER-3)
Aug 12 17:17:16 CEST: 0x00 0x00 0x00 0x00 0x03 0x07 0x06 0x06 (STACK_USER-3)
Aug 12 17:17:16 CEST: 0x00 0x20 0xAA 0x0F 0xFF 0x00 0x00 0x80 (STACK_USER-3)
Aug 12 17:17:16 CEST: 0x00 0x00 0x00 0x43 (STACK_USER-3)
Aug 12 17:17:16 CEST: (STACK_USER-3)
Aug 12 17:17:16 CEST: next reg: (STACK_USER-3)
Aug 12 17:17:16 CEST: 0x80 0xE4 0x00 0x00 0x00 0x00 0x52 0x00 (STACK_USER-3)
Aug 12 17:17:16 CEST: 0xC4 0x37 0x00 0x00 0x00 0x00 0x00 0x00 (STACK_USER-3)
Aug 12 17:17:16 CEST: 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 (STACK_USER-3)

What does 0x08 at which location exactly stand for? Do you have any table or tool for that?