Error Appearing in logs of a newly installed Cisco 3850 (WS-C3850-12S)

[netops2014]

Hi,

We recently installed two Cisco 3850 SW version 16.6.5, Image CAT3K_CAA-UNIVERSALK9, model WS-C3850-12S on our campus network. 

One in each of our data centres. These are connected together on a trunk link with SM fibre and 10G-LR SFPs.

We run HSRP between both data centres and we have noticed an error appearing on the standby router in the last couple of days. Our Active router is identical (model, SW Version and Image) is not showing the error. Our PRTG monitoring doesn't show and spikes or issues with memory, CPU or env. I fear this maybe a bug and am afraid of a reboot as these routers connect to our critical systems network. Load on the network isn't much which approx 1GB running over the 10G trunk and 400MB to the internet which is also a 10G link.

Has anyone seen this before by any chance?

Aug 13 01:35:54: %IOSXE-4-PLATFORM: Switch 1 R0/0: kernel: EDAC DEVICE1: CE: octeon_l2c_err instance: l2c0 block: l2c1 'L2D Single bit error (corrected):type:2, syn:0x241, way:42129'
Aug 13 01:40:12: %IOSXE-4-PLATFORM: Switch 1 R0/0: kernel: EDAC DEVICE1: CE: octeon_l2c_err instance: l2c0 block: l2c1 'L2D Single bit error (corrected):type:2, syn:0x241, way:42129'
Aug 13 01:42:51: %IOSXE-4-PLATFORM: Switch 1 R0/0: kernel: EDAC DEVICE1: CE: octeon_l2c_err instance: l2c0 block: l2c1 'VBF Single bit error (corrected):type:0, syn:0x241, way:42129'
Aug 13 01:47:23: %IOSXE-4-PLATFORM: Switch 1 R0/0: kernel: EDAC DEVICE1: CE: octeon_l2c_err instance: l2c0 block: l2c1 'L2D Single bit error (corrected):type:2, syn:0x241, way:42129'
Aug 13 01:51:53: %IOSXE-4-PLATFORM: Switch 1 R0/0: kernel: EDAC DEVICE1: CE: octeon_l2c_err instance: l2c0 block: l2c1 'L2D Single bit error (corrected):type:2, syn:0x241, way:42129'
Aug 13 01:51:54: %IOSXE-4-PLATFORM: Switch 1 R0/0: kernel: EDAC DEVICE1: CE: octeon_l2c_err instance: l2c0 block: l2c1 'L2D Single bit error (corrected):type:0, syn:0x241, way:42129'
Aug 13 01:51:54: %IOSXE-4-PLATFORM: Switch 1 R0/0: kernel: EDAC DEVICE1: CE: octeon_l2c_err instance: l2c0 block: l2c1 'VBF Single bit error (corrected):type:0, syn:0x241, way:42129'
Aug 13 01:52:39: %IOSXE-4-PLATFORM: Switch 1 R0/0: kernel: EDAC DEVICE1: CE: octeon_l2c_err instance: l2c0 block: l2c1 'L2D Single bit error (corrected):type:2, syn:0x241, way:42129'
Aug 13 01:52:40: %IOSXE-4-PLATFORM: Switch 1 R0/0: kernel: EDAC DEVICE1: CE: octeon_l2c_err instance: l2c0 block: l2c1 'L2D Single bit error (corrected):type:2, syn:0x241, way:42129'
Aug 13 01:59:04: %IOSXE-4-PLATFORM: Switch 1 R0/0: kernel: EDAC DEVICE1: CE: octeon_l2c_err instance: l2c0 block: l2c1 'L2D Single bit error (corrected):type:2, syn:0x241, way:42129'
Aug 13 02:04:22: %IOSXE-4-PLATFORM: Switch 1 R0/0: kernel: EDAC DEVICE1: CE: octeon_l2c_err instance: l2c0 block: l2c1 'L2D Single bit error (corrected):type:2, syn:0x241, way:42129'

Thanks

[marce1000]

 - That looks like a hardware error on first sight, but could be a software bug too. As the single bit error is apparently corrected I would leave the setup untouched (presuming the environment is production critical).

 M.

[Giuseppe Larosa]

Hello netops2014,

the error

>> Aug 13 01:47:23: %IOSXE-4-PLATFORM: Switch 1 R0/0: kernel: EDAC DEVICE1: CE: octeon_l2c_err instance: l2c0 block: l2c1 'L2D Single bit error (corrected):type:2, syn:0x241, way:42129'

It saying it had to change a single bit value in a l2c0 block, but it has been corrected.

It may be hardware related or a SW bug

I agree with Marce1000, for the moment if there is no evidence of impact on network, I would wait without taking any action.

Edit:

octeon is the name of a processor used in several Cisco products and linecards. Octeon is a multi core processor used also in Juniper SRX firewall if I correctly remember

For example I have found a bug mentioning the octeon processor.

https://bst.cloudapps.cisco.com/bugsearch/bug/CSCud03402

it is considered a cosmetic log with no impact. We can hope the same for your case.

There are other bugs more severe reporting of octeon crash or memory corruption (with no details).

Hope to help

Giuseppe

[Leo Laohoo]

Raise a TAC Case. It could be CSCur77024.

[netops2014]

Hi all,

Thanks for taking the time to take a look at this. The errors are still there but there has been no impact at all so far. I unfortunately don't have TAC on these boxes, so I will hold off on making any changes to the 3850. If anything happens or if there is a change with this, i will update you all. Thanks!

[Leo Laohoo]

Upgrade the firmware to 16.6.6 and don't leave 16.6.X train.