Forcing Link Down as Unrecoverable Error detected in chip/fw

[wilber822]

My HP blades were running very stable before I upgrade to  ESXi 5.5 along with firmware upgrading.

I'm using ESXi 5.5 U1 now, HP BL460c G6, G7 and Gen8 blades.

Hardware  firmware and drivers has been  upgraded from SPP2013.02 to SPP2014.06. Then I got random ESXi hosts lost network connectivity.

HP asked me upgrade to SPP2014.09. So the HP network adapter NC553i (Emulex) firmware is 10.2.340.19, driver is 10.2.298.5. Issue still present.

Whenever the issue present, I got following error in vmkernel. Only reset blade can fix the problem.

No error log in VC or IML except In Pause Frame significantly increased between the problem blade and VC down link.

Does anybody see same problem like me?

[MKguy]

There are a few people with the same issue if you look through the other thread you've posted to:

https://communities.vmware.com/message/2422395#2422395

Seems like it's a general issue with the new driver type for emulex-based CNAs (elxnet), which can be worked around by switching to the older driver (be2net):

esxcli system module set --enabled=false --module=elxnet

esxcli system module set --enabled=true --module=be2net

Reboot.

Have you tried this?

Also this KB article mentions there is a driver newer than 10.2.298.5 available from Emulex (not public) which fixes some issues:

VMware KB: Packet drops and connectivity issues when using Emulex elxnet Driver version 10.2.298.5 o...

[wilber822]

Hi MKguy

Thanks a lot for the information. Yes, I had read the thread you provided before posting.

Since the system is production, I really want to apply official solution.

The secondary KB is not exactly same like my case, but I will take a look release notes.

[wilber822]

Hi MKguy

Per-talked with HP, this issue not only observed on ESXi but also observed on Linux blade.

No fix at this moment, I'm still waiting HP and VMware feedback....

The issue looks like occurred after Emulex changed their version family from 4.9.x to 10.2.x

[vMarkusK1985]

Hi,

is there a newer update to this issue avaiable?