Hi, I have a dell 1950 with a Broadcom/Qlogic BCM5708 nic running ESXi 5.5 that disconnects after a few hours. If I shut the interface on the ESXi side and bring it back up then it starts working again. This is a 2 port Card but I see the problem only with 1 of the Nic's.
~ # ethtool -i vmnic0
driver: bnx2
version: 2.2.5f.v55.8
firmware-version: 7.8.53 bc 7.4.0 UMP 1.1.9
bus-info: 0000:05:00.0
~ #
I have updated the drivers and firmware but it has not helped. The old driver and firmware were
driver: bnx2
version: 2.2.5f.v55.6
firmware-version: bc 4.4.1 UMP 1.1.9
# esxcli network nic list
Name PCI Device Driver Link Speed Duplex MAC Address MTU Description
------ ------------- ------ ---- ----- ------ ----------------- ---- -------------------------------------------------------------
vmnic0 0000:005:00.0 bnx2 Up 1000 Full 00:19:b9:e6:91:d6 1500 Broadcom Corporation Broadcom NetXtreme II BCM5708 1000Base-T
vmnic1 0000:009:00.0 bnx2 Up 1000 Full 00:19:b9:e6:91:d8 1500 Broadcom Corporation Broadcom NetXtreme II BCM5708 1000Base-T
#
I have gone through quite a few links related to the BCM5708 and one of the links suggested I turn of MSI-X in vmware.
The link was kb.vmware.com/.../search.do;cmd=displayKC&externalId=1029070
I have since reverted back because it did not help and I also read somewhere that MSI-X does not apply to the BCM5708.
There were some other articles related to the BCM5708 and changing the Rx Buffer ring size value as per this URL "support.qlogic.com/.../2756;popup=true". The above setting did not help either. I went as high as 4080. The default was 255.
The weird thing is that I have another server with the exact same hardware specs but it runs an older firmware and driver which I cannot seem to find and that's working fine. Just for reference the other working server (separate physical server) has the following firmware
/var/log # ethtool -i vmnic1
driver: bnx2
version: 2.2.5f.v55.6
firmware-version: bc 2.9.1
bus-info: 0000:09:00.0
Strangely the working server does not mention anything about UMP, which I believe is Universal Management Port, so am wondering if that is what is causing issues because even after updating firmware on the non-working server the UMP version has not changed. I believe BMC & IPMI are disabled on the non-working server but regardless the settings for that portion are similar across both devices except for the firmware version of the BMC which is lower on the working device. These are the settings for BMC on the non-working device
 |  | |  | | ||||||||||||||||||||||
|
Can someone please advise what the issue might be and why is it that every few hours the VM associated with this vmnic0 get disconnected. vmnic1 on this same non-working server appears to be running absolutely fine without any issues. Thanks