Intel Pro/1000 PT Quad PCIe doesn't work...

[jbsengineer]

Just an FYI. There were some threads about this but I couldn't find them.

We got some new Dell 2950's in with the PCIe architecture. With Intel having the new Quad PCIe NIC we decided to wing it and see if it wouldn't atleast work in our test labs. From what I've read it's the same chipset. However it doesn't work. ESX won't load a driver for it.

So my question is how do I let VMware know that atleast I would love to see it supported in the next release?

[jasper9890]

Greetings everyone.. I'm a total newbie to VMware so pardon the inexperience. I've been following this thread as my brand new systems (that were speced by Dell VMware engineers i might add) are having this problem. Am i really to believe that up until this point quad port cards have never worked? Just the intel one?

Thanks..

[robbiehe]

Lambeth: any news about the patch release?]

Thanks

[ErMaC1]

There's a quad-port PCI-X Intel NIC that we use in all our servers (HP DL385 G1's) and it works great. The problem is the PCI-e version of this card isn't supported by VMWare yet - the problem isn't with the fact that it's quad-port, per se.

[amickiewicz]

Check out the VMware patch released yesterday:

http://www.vmware.com/download/vi/esx-1006511-patch.html

[ErMaC1]

Is there a URL to see when VMWare releases new downloads like this, or other hotfixes?

[amickiewicz]

http://www.vmware.com/download/

Just click on your ESX version to look for any updates/patches.

If you google for your issues more often then not you will find a vmware kb article with a link to a patch or workaround to resolve the issue.

[robbiehe]

yeahhhhh everything is working now perfectly!! the ESX-1410076.tgz patch is working flawlessly....

Tnx

[jasper9890]

Same here! Does anyone know what mailing list or RSS feed i can find to get announcements on patches?

[jatsma]

Hello again.

We downloaded the patch ESX-1006511 and everything works fine now.

Thanks for releasing the patch.

[RobvanderHelm]

Hi, we are using the patched drivers released from Vmware for the Intel Quad PT card but we have another problem.

When we patch one lan interface, two connections in Virtual Center report a active link.... This occurs on two identical XBM x3850-8864 servers with two Intel Pro1000PT Quad cards each. Virtual machines on the different hosts cannot see each other. It looks like the network cards do not function correctly.

anyone an idea?

Message was edited by:

RobvanderHelm

[wbeede]

I am having this same issue on the x3850-8864 with the Intel Pro1000PT Quad card, one port plugged in and two show online. Was this ever resolved? I have an open ticket with VMWare and IBM, both say the card is supported but can't seem to help.

[phanman]

I am also running ESX 3.01 with 3 IBM x3850-8864 and am experiencing very strange issues with the Intel Pro1000PT PCIe 8x adapters.

I'll pull the physical cable from one of the nics and the vmnic will still appear as connected and i'll be able to get an ip address. I'm also trying to use seperate ports on the card for different vlans and it'll show all adapters as being on one vlan.

When I vmotion to another ESX both within the cluster (all vswitch configs are identical) I'll have to disconnect the nic under edit settings and reconnect just to get connectivity almost every time. (Very annoying)

I thought maybe this was an issue with our HP procurve switches until I plugged a simple 10/100 switch in between them and noticed that the autonegotiate speed still claimed to be 1000/FULL. (Even after cycling the options)

Any help or light on this situation would be much appreciated. That is only a small list of quirky issues with these nics.

PS. we already have ESX-1006511 patch applied.

Message was edited by:

phanman

[Candi]

I ran into this problem, and was told by vmware support that only specific revs of the card were supported (I was told this only after working with support and yelling at my hardware reseller for days). Too much trouble to deal with and I had the slots to spare, so I swapped in dual ports instead.

[fredz]

I ran into this problem, and was told by vmware

support that only specific revs of the card were

supported (I was told this only after working with

support and yelling at my hardware reseller for

days).

Anyone know which revs are (not) supported?

[mtsm]

im getting crazy....i used the patch , no success ....the weird is that it was running before

it's an Intel PRO/1000 MT server adapter (single) ....changed the cards and same thing!! here it hoes

---

Mar 7 02:15:26 s6004vm12 vmkernel: 0:02:17:29.625 cpu1:1034)<6>Intel(R) PRO/1000 Network Driver - version 7.0.33
Mar 7 02:15:26 s6004vm12 vmkernel: 0:02:17:29.625 cpu1:1034)<6>Copyright (c) 1999-2005 Intel Corporation.
Mar 7 02:15:26 s6004vm12 vmkernel: 0:02:17:29.625 cpu1:1034)PCI: driver e1000 is looking for devices
Mar 7 02:15:26 s6004vm12 vmkernel: 0:02:17:29.625 cpu1:1034)PCI: Trying 01:01.0
Mar 7 02:15:26 s6004vm12 vmkernel: 0:02:17:29.625 cpu1:1034)PCI: Trying 01:01.1
Mar 7 02:15:26 s6004vm12 vmkernel: 0:02:17:29.625 cpu1:1034)PCI: Trying 01:02.0
Mar 7 02:15:26 s6004vm12 vmkernel: 0:02:17:29.625 cpu1:1034)PCI: Trying 02:01.0
Mar 7 02:15:26 s6004vm12 vmkernel: 0:02:17:29.625 cpu1:1034)PCI: Announcing 02:01.0
Mar 7 02:15:26 s6004vm12 vmkernel: 0:02:17:29.625 cpu1:1034)PCI: Trying 04:01.0
Mar 7 02:15:26 s6004vm12 vmkernel: 0:02:17:29.625 cpu1:1034)PCI: Announcing 04:01.0
Mar 7 02:15:26 s6004vm12 vmkernel: 0:02:17:29.625 cpu1:1034)PCI: Trying 0a:01.0
Mar 7 02:15:26 s6004vm12 vmkernel: 0:02:17:29.625 cpu1:1034)PCI: Announcing 0a:01.0
Mar 7 02:15:26 s6004vm12 vmkernel: 0:02:17:29.855 cpu1:1034)<3>e1000: 0a:01.0: e1000_probe: The EEPROM Checksum Is Not Valid
Mar 7 02:15:26 s6004vm12 vmkernel: 0:02:17:29.855 cpu1:1034)PCI: Trying 0c:01.0
Mar 7 02:15:26 s6004vm12 vmkernel: 0:02:17:29.855 cpu1:1034)PCI: Announcing 0c:01.0
Mar 7 02:15:26 s6004vm12 vmkernel: 0:02:17:30.086 cpu1:1034)<3>e1000: 0c:01.0: e1000_probe: The EEPROM Checksum Is Not Valid

Mar 7 02:15:26 s6004vm12 vmkernel: 0:02:17:30.086 cpu1:1034)PCI: driver e1000 claimed 0 device

Mar 7 02:15:27 s6004vm12 vmkernel: 0:02:17:30.086 cpu1:1034)ALERT: Mod: 490: Initialization for e1000 failed with -19.

Mar 7 02:15:27 s6004vm12 vmkernel: 0:02:17:30.086 cpu1:1034)Module failed to load with status -19.

Mar 7 02:15:27 s6004vm12 vmkernel:

Mar 7 02:15:27 s6004vm12 vmkernel: 0:02:17:30.086 cpu1:1034)Mod: 501: mainHeap avail after: 13186024

any ideas?

the support didnt help me shit

[mtsm]

ok after some research i found out some interesting stuff...this problem seems to be related to power savings in pci-x slot..

i booted the machine into the ubuntu live cd....had the same message...

but the e1000 driver at live CD has an option (eeprom_bad_csum_allow=1) which will ignore the eepprom checksum...the drivers loads and the cards works okay...where can i find vmware's e1000 driver source code so i can recompile and make it ignore this checksum ....

any ideas?

[norx]

I am about to install two serveres with 2 x Intel PRO/1000 PT Quad Port in each server.

For me the driver is found, and nic is listed in GUI.

There is Link on switch and on network adapter, but inside VMware it shows as link is down.

[Candi]

Norx-

Which port on the card are you plugged in to? The issue I was seeing was the top port registered on ports 1 and 3 within the gui, and the second port registered on ports 2 and 4. Anything plugged into 3 or 4 did nothing.

Is this what you are seeing?

-Candi

[norx]

Hi Candi,

I have not been out at the customer location for a while so I have not been checking all the combinations, but If all cabeles are pluged inn I have link in GUI on all ports, switch and back of Network adapter.

If I unplug one cable, "top port", all of the ports is marked as Down in GUI, but link is shown on switch and network adapter.

Different behavior in GUI if I leave Only one network cable on the adapter.

So it sounds like I have the same issue as you got...

[norx]

I Finaly found the error. Card is now working on the IBM Server.

Firmware upgrade the bios on the IBM to 1.04. If you run the express update you will only update to 1.03. The 1.04 version fixes some PCI Bridge errors.

Message was edited by:

norx