VMware Cloud Community
kinsden
Enthusiast
Enthusiast

VMKernel NIC

Hi all,

OK..here's the situation...we were having an ESX 3.5 host on a HP ProLiant DL580 G5. We swapped the disks to a new HP DL580G5 (as the former one was on loan). Since the change, I am not able to make the VMKernel portgroup work. Whatever NIC I attach to the vswitch that has VMKernel portgroup, its link does not come up or if it comes up, goes down and then comes back up. Its flapping!

The physical switch port is 100MB Full and I have configured the NIC to the same speed/duplex setting. But it still shows as down in VI Client and service console. The NICs are Intel PRO/1000 PT Low Profile Quad Port PCIe Gigabit Ethernet Controller.

No matter which port I use on the physical NIC or which physical NIC I use, it doesn't seem to like it. Also I tried setting speed/duplex setting using esxcfg-nics command from SC, but it does not seem to taking any effect.

Any help would be greatly appreciated!! Many thanks.

Tags (3)
Reply
0 Kudos
6 Replies
admin
Immortal
Immortal

Have you see something on the vmkernel log (/var/log/vmkernel)?

Reply
0 Kudos
kinsden
Enthusiast
Enthusiast

Its full of repeating entries of:

vmkernel:0:03:20..........e1000: vmnic10: e1000_watchdog:NIC Link is Down

vmkernel:0:03:20..........e1000: vmnic10: e1000_watchdog:NIC Link is Up 100 Mbps Full Duplex

vmkernel:0:03:20..........e1000: vmnic10: e1000_watchdog:10/100 speed: disabling TSO

Reply
0 Kudos
admin
Immortal
Immortal

Did you enable TSO on the BIOS of your server? Is is this functionnality was enable on your last server&

Reply
0 Kudos
kinsden
Enthusiast
Enthusiast

I haven't enabled TSO and sorry, but I am not sure about my last server (that's why I am (trouble)shooting in dark) Smiley Sad

Reply
0 Kudos
MoparMark
Contributor
Contributor

Any head way on this problem? I'm experiencing the same problems and have found references to this issue all over the net, but never any solutions. System is an Dell Optiplex 755, Core2Duo, 4 gigs of RAM. I'm using on the onboard gig Intel NIC hooked up to a Cisco 3500 series switch. VM Server is running on a stripped down install of Ubuntu x64 8.04. Keeping getting the...

eth2: Link is Down

eth2: Link is Up 100 Mbps Full Duplex, Flow Control: RX

eth2: 10/100 speed: disabling TSO

... over and over again in dmesg. It's really starting to interfere with the usage of VM machines. Any help would very much be appreciated.

Reply
0 Kudos
RedLimey
Contributor
Contributor

Hi - I just posted on another thread asking if there has been any progress in people finding a solution. Something of interest here (almost 2 years after you posted!) - you quoted you are using Cisco 3500's. I am ignorant when it comes to network devices and their popularity (are 3500's extremely common place - the standard for Cisco shops?), but we have about 7 or 8 remote hosts (out of ~70) with PCI Adapters that exhibit the link up/down issue (at varying interval). I've had it confirmed that we are using 3500's for the most part. We use HP Proliant DL385's, and the integrated NICs do NOT have the link problem, just the added PCI adapters.

We are also a manufacturing company. Perhaps a dumb question/thought, but I wonder if it's a noise/interference issue and a combination of sensitive NICs/unforgiving switch models.

Reply
0 Kudos