UCS 5108 - CRCDelta Error

[Quintin.Mayo]

Hi,

We are seeing the following error " TCA: etherNiErrStats crcDelta for sys/chassis-2/slot-2/fabric/port-1 , current-value = 5, raised above esc value 5" on 2 of our UCS chassis's.  The system details are FI System is 5.0(3)N2)3.12E , FI Model is UCS6284UP, UCS Version 3.1.(2e) and UCS-IOM-2208XP.  Can someone confirm this is related to a bug? Any assistance would be greatly appreciated.

Thanks,

[Wes Austin]

Its not a bug, its telling you that you have incrementing CRC errors on the FI. You can run the following commands from NXOS to check where they are coming from. Run this from A and B side and see where you are incrementing CRC.

UCS#connect nxos a|b

UCS-B(nxos)#show interface counters errors

[Quintin.Mayo]

Hi,

We implemented the commands as directed and found the below interfaces on side A and B.  Is it a matter of bouncing the interfaces and/or possibly changing the cabling? Please inform of the resolution for the issue. Once again your assistance is greatly appreciated.

Thanks,

[Quintin.Mayo]

The interfaces are below
B side
Eth1/25 0 503887 0 503887 0 0
Eth1/26 0 503806 0 503806 0 0
Eth1/27 0 10 0 10 0 0
Eth1/28 0 0 0 0 0 0
Po2 0 1007703 0 1007703 0 0

A side
Eth1/25 0 502391 0 502391 0 0
Eth1/26 0 502310 0 502310 0 0
Eth1/27 0 0 0 0 0 0
Eth1/28 0 0 0 0 0 0
Po1 0 1004701 0 1004701 0 0

[Wes Austin]

Since those are across all of your interfaces, you likely have frames traversing the UCS fabric that are >1500 MTU. Have you made any recent changes in the network or elsewhere to use jumbo frames? 9000 MTU?

If you can track down what hosts are sending or receiving jumbo MTU you can clear the errors. You alternatively may increase your QoS policy in UCSM to allow 9000 MTU for the best effort class in order to clear the errors as well.

[Quintin.Mayo]

Hi,
Thanks for your assistance. Can you provide documentation on how to do the procedure needed for clearing the errors and configuring the best effort class for 9000 MTU. Once again you are truly appreciated.

Thanks,

[Wes Austin]

You can clear the errors via SSH to the UCS. Same location as the commands you ran previously, you just run 'clear counters'.

This document outlines how to change the best effort class in UCSM to 9216 MTU (best practice). You do not need to adjust it anywhere else, unless you have vNICs that you also want to allow to utilize >1500 MTU.

https://www.cisco.com/c/en/us/support/docs/servers-unified-computing/ucs-b-series-blade-servers/117601-configure-UCS-00.html 

[Quintin.Mayo]

Hi,

If we increase the MTU size to 9000 to allow for jumbo frames will this cause any disruptions to our current environment as we don't allow Jumbo frames at this time.  Any assistance will be greatly appreciated.

Thanks,

[Wes Austin]

You should be adjusting to 9216 as best practice. Yes, this can cause disruption and recommend adjusting this during maintenance window.

https://www.cisco.com/c/en/us/td/docs/unified_computing/ucs/ucs-manager/GUI-User-Guides/Network-Mgmt/3-2/b_UCSM_Network_Mgmt_Guide_3_2/b_UCSM_Network_Mgmt_Guide_3_2_chapter_0111.html