Hi Guys,
I've recently changed my server from a ubuntu installation with vmware server to a esxi installation. On my server i have 2 network cards (see lspci below), 1 intel 100mbit and a intel 1000mbit adapter.
If i send (a large amount of) data to the host (using scp, or wget) the gigabit adapter disconnects and reconnects to the network (i can see it on my switch) and in my /var/log/messages (when transfering to the 100mbit adapter i dont have this issue)
/var/log/messages
May 10 12:41:19 vmkernel: 1:16:27:04.617 cpu0:1029)<3>e1000: vmnic0: e1000_tx_hang_debug: buffer_info[66] informations
May 10 12:41:19 vmkernel: skb <0x0>
May 10 12:41:19 vmkernel: dma <0x0>
May 10 12:41:19 vmkernel: 1:16:27:04.617 cpu0:1029)<3>e1000: vmnic0: e1000_tx_hang_debug: time_stamp <14562169>
May 10 12:41:19 vmkernel: length <0>
May 10 12:41:19 vmkernel: next_to_watch <67>
May 10 12:41:19 vmkernel: 1:16:27:04.617 cpu0:1029)<3>e1000: vmnic0: e1000_tx_hang_debug: tx_ring 66 descriptor informations
May 10 12:41:19 vmkernel: tx_desc.lower.length <0>
May 10 12:41:19 vmkernel: tx_desc.lower.cso <0x0>
May 10 12:41:19 vmkernel: tx_desc.lower.cmd <0x21>
May 10 12:41:19 vmkernel: 1:16:27:04.617 cpu0:1029)<3>e1000: vmnic0: e1000_tx_hang_debug: tx_desc.upper.status <0x0>
May 10 12:41:19 vmkernel: tx_desc.upper.css <0x0>
May 10 12:41:19 vmkernel: tx_desc.upper.special <0x0>
May 10 12:41:19 vmkernel: buffer_addr <0x322200000000>
May 10 12:41:19 vmkernel: 1:16:27:04.617 cpu0:1029)<3>e1000: vmnic0: e1000_tx_hang_debug: buffer_info[67] informations
May 10 12:41:19 vmkernel: skb <0x3ea042c8>
May 10 12:41:19 vmkernel: dma <0x40404430>
May 10 12:41:19 vmkernel: 1:16:27:04.617 cpu0:1029)<3>e1000: vmnic0: e1000_tx_hang_debug: time_stamp <14562169>
May 10 12:41:19 vmkernel: length <114>
May 10 12:41:19 vmkernel: next_to_watch <67>
LSPCI:
06:01.00 Network controller: Intel Corporation 82541PI Gigabit Ethernet Controller
06:08.00 Network controller: Intel Corporation 825XX Ethernet Adapter
I've tried 1000mbit full, 100mbit full, 100mbit half, 10mbit full duplex. all these speeds are causing problems, only the 10mbit half duplex seems to be working
Has anybody seen this problem before? or can somebody point me in the correct direction? All your thoughts are welcome and i'm happy to provide more information if needed.
Thanks!
Henk
As a test i created a guest and copied some files to that host.. same problem, network also disconnects
Maybe a bad Ethernet cable? Or bad port on your switch?
What version of ESXi do you have installed? I think build 158874 is the latest available.
i've tested it with different cable, attached directly to my laptop, same problem..
ps. running version: VMware ESX Server 3i U4 Installable Refresh Version 3.5 Update 4 | 153875 - 03/30/09
Porblem look like faulty nic or duplex miss match please check device driver. also one thing try to ping with packet size to check link load.
- If you found this information useful, please consider awarding points for "Correct" or "Helpful". Thanks!!!
Thanks
Satish Patel
Journey toward "Virtual world"
Host that works fine
--- 192.168.60.43 ping statistics ---
74100 packets transmitted, 74100 received, 0% packet loss, time 18389ms
rtt min/avg/max/mdev = 0.085/0.191/4.052/0.042 ms, ipg/ewma 0.248/0.188 ms
Host that fails:
--- 192.168.60.40 ping statistics ---
11085 packets transmitted, 8608 received, 22% packet loss, time 32576ms
rtt min/avg/max/mdev = 0.215/0.246/3.854/0.138 ms, ipg/ewma 2.939/0.239 ms
ps. running version: VMware ESX Server 3i U4 Installable Refresh Version 3.5 Update 4 | 153875 - 03/30/09
Please try installing the latest updates for ESX3i -- The update I am running is build 158874. The easiest way to get/install updates is by running the VMware Infrastructure Update program from the same Windows machine that you installed the VMware Infrastructure Client program to (provided that you answered yes to the question about installing the Update service during the install of your VI Client)
06:01.00 Network controller: Intel Corporation 82541PI Gigabit Ethernet Controller vmnic0
FYI, I have that same NIC installed in my white box ESX3i server w/out problems. Try moving it to a different slot? But I'd try installing the latest version of ESXi first.
Host that works fine
--- 192.168.60.43 ping statistics ---
...
Host that fails:
--- 192.168.60.40 ping statistics ---
11085 packets transmitted, 8608 received, 22% packet loss, time 32576ms
rtt min/avg/max/mdev = 0.215/0.246/3.854/0.138 ms, ipg/ewma 2.939/0.239 ms
When you say "Host" above do you actually mean Guest Virtual Machines? Could you have a MAC address conflict between one or more of your Guests (sometimes happens if a Guest is manually cloned)?
I have exactly the same problem - I've tried two different switches (cables are fine). My box is a Dell PowerEdge 750.
The problem also means I cannot send the update to the server as it disconnects from the LAN at 11% of the transfer - the only way to get it back is to restart the managment interface.
Hi Guys,
Today i went to the datacenter with a new nic. i replaced the adapter but the same problem occured with the new card.
after fiddeling around a bit, i found the following sollution: in the original situation the network adapter was in a pci-riser at slot: pciB, on slot pciA a sata raid controller was attached. As i didnt need the sata controller anymore, i removed it, and moved the nic from pciB to pciA. (i assumed pciA was the best and shortest path.
but pciA is connected true a little cable, while pciB is connected directly to a pci slot. moving the network adapter back to pciB resolved all of my problems
Thanks for your help!