VMware Cloud Community
jedijeff
Enthusiast
Enthusiast

Networking Bug in KB 2072694

I have a discussion in ESXi, but thought I would post my followup in the networking forum,which is where I should have posted the original question I suppose:

I got some more information on this bug, in that the fix that was in the KB that I referenced, was how to fix it when the you have updated and it you have the problem. You may experience the issue if you move to Update 2 and keep the TSO on in E1000 on the host, and then vmotion a vm with e1000 this vm may lockup and lose networking.

If you want to deploy U2, you can simply make the changes before any VMs move to the host.

Now, does anyone know of any issue if you have a guest(rhel 5.5) with e1000 that is using TSO, but the host has it turned off?

Also this will be fixed before U3.

Reply
0 Kudos
4 Replies
jedijeff
Enthusiast
Enthusiast

After thinking about this more, I can't see any issue with TSO off on the host, and on on the guest. It would be just like a physical box with a crappy Nic that doesnt do TSO, yet the driver has it on by default. Thats what I think anyway.

Reply
0 Kudos
HeathReynolds
Enthusiast
Enthusiast

Somewhat related to this issue - Here is a one liner PowerCLI to dump a list of all of your E1000 nics.

Get-vm | get-networkadapter | where-object {$_.type -like "*e1000*" } | Export-Csv c:\Frain\network.csv -NoTypeInformation

I have a lot more e1000s that I thought I did.

My sometimes relevant blog on data center networking and virtualization : http://www.heathreynolds.com
MKguy
Virtuoso
Virtuoso

The issue has now been resolved in a new patch released a few days ago:

VMware KB:    VMware ESXi 5.1, Patch ESXi510-201402401-BG: Updates esx-base   

This patch updates the esx-base VIB to resolve the issue where virtual machines using the E1000/E1000E adapter might lose network connectivity when the networking heap memory, netGPheap, reaches the maximum limit. In the vmkernel.log file, you see entries similar to:

WARNING: Heap: 2677: Heap netGPHeap already at its maximum size. Cannot expand.

WARNING: Heap: 3058: Heap_Align(netGPHeap, 35156/35156 bytes, 8 align) failed. caller: 0x41800915066c

WARNING: E1000: 8817: failed to enable port 0x200002f on DvsPortset-0: Out of memory

WARNING: Net: vm 367615: 4454: cannot enable port 0x200002f: Out of memory

For more information about symptoms and resolution, see KB 2072694.

-- http://alpacapowered.wordpress.com
Reply
0 Kudos
jedijeff
Enthusiast
Enthusiast

Yeah I saw that earlier. Quick fix Smiley Happy Good job for vmware. I was already planning rolling out some TSO  options, this saves me tons of work.

Reply
0 Kudos