Cisco 3850 Redundant Power Supply Reporting Issue

[mchirico01]

We have deployed several Cisco 3850's. Several are showing the following symptom

ont2-tc3-es01#sh env power
SW  PID                 Serial#     Status           Sys Pwr  PoE Pwr  Watts
--  ------------------  ----------  ---------------  -------  -------  -----
1A  PWR-C1-715WAC       DCB1702G0QU  OK              Good     Good     715
1B  PWR-C1-715WAC       DCB1702G0G6  No Response     Good     Good     715

Every single switch at this particular location has the same status on the B power supply. The problem was there on the previous code level (3.2.2) and we recently upgraded to 3.3.5 to resolve some SNMP bugs on the switches we also expected the latest Assurewave version to fix this issue also. So couple questions:

1. Anyone else seeing this?

2. Anyone found a fix for it?

               Thanks!

                          Marty

[Petr Nyvlt]

I have the same issue, only sw version is 03.03.01SE.

It alse reports this undocumented error into log:

%NGWC_PLATFORM_FEP-1-FRU_PS_ACCESS: Switch 1: power supply B is not responding

[mchirico01]

I got a ticket open with TAC, per the engineer there are two bugs that solved this issue. One was resolved in 3.3.5, the other he thinks is resolved in the 3.3.7 - but he is going to confirm with the developers.

He mentioned that reinserting the power supply might force it correct the error, but unfortunately my cabinets are about 30Ft up in the air (factory warehouse environment) so I can't confirm

I'm holding off on doing anymore OS upgrades until I get confirmation that it's fixed.

 I'll keep the thread updated

               thanks!

                         Marty

[zabeelmusa]

Hi Marty

Did you get an update from TAC on this issue.  A customer of ours is experiencing the same issue on 3850s running 03.02.03.

Bug search shows this as fixed on CSCue44402

Regards

Zabeel

[mchirico01]

Hi Zabeel,

                   I did talk with TAC - per the engineer there are two bugs that cause the issue and one was fixed in 3.3.5 - so we upgraded all our 3850's to that code level. Unfortunately for us, that still didn't fix the issue. TAC suggested that it may be fixed in rev 3.7.0, but that has not officially been confirmed by Cisco's developers.

         We're waiting for the official "fix" before going through another round of switch upgrades.

        TAC did suggest reseating the power supply if you still have the issue after the upgrade to 3.3.5 - but that's not really an option for me since my switches are geographically spread out.

   HTH

             Marty

[zabeelmusa]

Thanks for the quick reply Marty much appreciated.

For the last 3 issues i had the PSU replaced only to find another switch today with the same issue which led me to doing bug search and this post.

The customer is on a old version so I will upgrade to the latest code and hope that it gets fixed.  I am in the same boat as the customer is remote and not an option to keep upgrading the firmware.

Regards

Zabeel.

[Suresh Gurugubilli]

on which IOS this bug fixed.

Cisco Bug: CSCuv18572

[gregorczyk]

Do you have confirmed good software for this switches ?

Is it rev 3.7.0 resolved issue or not ?

Waiting for update.

[mchirico01]

Hi Gregor,

              We haven't upgraded yet - but I can confirm reseating the non-responding power supply did fix the issue on the older code

[felixngu0260]

I encountered the same issue. Reseating the PSU resolved the issue but will need to monitor if it recurs. The switch is running on 03.02.03.SE

[Adam Swindell]

I am having this problem with about a dozen or so 3850's on 3.7.1. So I'd say 3.7.0 does not fix it. 

[0dmackessack]

Ok---we have had this problem on about 4 of these new 3850's---all at various locations and heights above various production lines---the 3850's continue to work well and a deeper analysis shows that both psu's are working well and out putting a good supply of power but falsly indicating no response---we opened a TAC case and were told that a future IOS upgrade would fix the problem but meanwhile this false error was causing havoc to our network management tools--we fixed all 4 by a combination of replacing the power cords or pulling the power supplies from their mains sockets---both fixes obviously ONE AT A TIME