%HARDWARE-3-MCU_I2C_READ_ERROR: I2C read failed in MCU and attached devices

[ayazali_ned]

Hi,

We hit the below mentioned issue with the following switch model : WS-C3750G-24PS"

"The switch is not detected the PoE devices on the ports and unable to provide power to the IP phones". Show power inline status shows off for most the IP Phones.

Gi1/0/6   auto   off        0.0     n/a                 n/a   30.0
 

Following log message, we had received on the switch when this issue was triggered. 

002652: 15w3d: %HARDWARE-3-MCU_I2C_READ_ERROR: I2C read failed in MCU and attached devices (e.g. Temperature Sensor) can't be read.

The issue is fixed after the switch IOS is upgraded to 12.2(55)SE9  C3750-IPBASEK9-M which is recently released on 18^th March 2014. Please confirm about IOS release 12.2(55) SE9 stability for the above mentioned error reported in 12.2(58)SE2. Is it a known bug or hardware issue for the particular platforms.

Regards,

Ayaz

[DEENA SLOAN]

I am also receiving this error and the WS-C3560G-48PS is running Version 15.0(2)SE9. 

Following log message, we received on the switch when this issue was triggered. 

Event : %HARDWARE-3-MCU_I2C_READ_ERROR: I2C read failed in MCU and attached devices (e.g. Temperature Sensor) can't be read..

From Cisco 3 years ago: The issue was fixed after the switch IOS is upgraded to 12.2(55)SE9  C3750-IPBASEK9-M which is recently released on 18^th March 2014. Please confirm about IOS release 12.2(55) SE9 stability for the above mentioned error reported in 12.2(58)SE2. Is it a known bug or hardware issue for the particular platforms.

Seems that this bug has reappeared and has not been fixed or returned  in 2017.

[Herbert Patrman]

Hello!

We're currently running 12.2(55r)SE on a WS-C2960S-48LPS-L and also receiving this message:

Aug 9 03:11:48 xxxxxxxx 54: Aug 9 01:12:01.573: %HARDWARE-3-MCU_I2C_READ_ERROR: I2C read failed in MCU and attached devices (e.g. Temperature Sensor) can't be read.
Aug 9 03:11:48 xxxxxxxx 55: Aug 9 01:12:02.574: %HARDWARE-3-MCU_I2C_BAD_DEV: I2C bad device address: 4 (alchemy:0)

As a consequent the switch turned of all power feeds to POE devices (12 IP cams class 3 with 15,4W per device)!

Further the switch started writing the following messages about 4hrs later (in between nothing!):

Aug 9 07:30:19 xxxxxxxx 56: Aug 9 05:30:32.069: %FRNTEND_CTRLR-1-MGR_TXQ_FULL: The front end controller Tx queue reached watermark level
Aug 9 07:35:20 xxxxxxxx 57: Aug 9 05:35:32.910: %FRNTEND_CTRLR-1-MGR_TXQ_FULL: The front end controller Tx queue reached watermark level
Aug 9 07:40:21 xxxxxxxx 58: Aug 9 05:40:33.761: %FRNTEND_CTRLR-1-MGR_TXQ_FULL: The front end controller Tx queue reached watermark level
Aug 9 07:45:21 xxxxxxxx 59: Aug 9 05:45:33.788: %FRNTEND_CTRLR-1-MGR_TXQ_FULL: The front end controller Tx queue reached watermark level
Aug 9 07:50:21 xxxxxxxx 60: Aug 9 05:50:34.545: %FRNTEND_CTRLR-1-MGR_TXQ_FULL: The front end controller Tx queue reached watermark level
Aug 9 07:55:22 xxxxxxxx 61: Aug 9 05:55:35.196: %FRNTEND_CTRLR-1-MGR_TXQ_FULL: The front end controller Tx queue reached watermark level
Aug 9 08:00:22 xxxxxxxx 62: Aug 9 06:00:35.286: %FRNTEND_CTRLR-1-MGR_TXQ_FULL: The front end controller Tx queue reached watermark level

...

until Aug 13 11:57:52 - then it was rebooted for changing power cord.

Some additional informations: this actual switch already has been replaced by RMA about 4 weeks ago because the old one suddenly began to shutdown itself completly. The new one also shows the same symptoms after some days. Powercord has been replaced and is pluged together with an other switch which doesn't show this problem.

The POE devices now have been attached to power injectors to get sure that this is not the trigger.

So is this a hardware problem caused by one or more PE devices or the switch itself or is it a software problem indeed?

Best regards,

Herbert