We've started having a problem with one of our clusters. One of the servers has decided it no longer wants to see the network over the x520 T2 card. The work around seems straight forward - simply disable the adapter and enable the adapter and it begins transmitting packets again. So not ideally practical when the server reboots as we start up with no networking.

The server is a Dell PowerVault NX3000 (2950) in a Windows 2008 R2 Enterprise cluster.

It uses it's Broadcom cards teamed with BACS for accessing the iSCSI SAN and the Intel X520-T2 as an ANS ALB Team to access the LAN. It's the LAN connectivity we lose until we do a disable/enable.

It all began following some windows updates, but oddly the other server in the cluster applied the same updates without this issue.

To try to resolve it I uninstalled using ProWin64 and made sure the adapters were gone from device manager. Then reinstalled. Still the same issue.

Can anyone help?

Ive got the Intel DQ67EP board with the Intel 82579LM onboard NIC.  Ive got the machine running Windows7 64bit.  Ive got this problem where randomly it drops from Gigabit to 10Mbps and will never negoiate back to gigabit.  I can disconnect the cable, reconnect it, dont work.  I can plug that same cable into another machine like my Lenovo T420 laptop, and it has no issues connecting gigabit, so I know its not the cable, or port, or Hub.  However still plugged into a different port of the hub, still dont work.  Plugged it into a different hub, still dont work, still keeps sitting at 10Mbps.

Ive rebooted it, still negoiates at 10Mbps.  Shut the PC down, wait 1min, turn it on, still sitting at 10Mbps.  So at this point dont believe its driver.  I have the latest 16.5 running on the Windows.

ONLY way I can get this working is one of two ways.  I have to unplug power, not just shut it down.  I need to kill power from the board entirely.  Wait 10secs, then plug it back in.  Then upon starting up, its set to gigabit with no issues.  Other way is to change a setting in Bios which when you exit/save it causes it to kill power to the board for a second.  So the fact that killing power to the board fixes it, tells me its some hardware or firmware level issue, clearly not the driver.

Now to explain when it happens, I do have the machine set to sleep and wake at certain times.  Most of the time it sleeps fine, wakes up fine and still sits at Gigabit.  Theres no explanation on how many times it takes before it decides to go into some stuck state of only negoiating at 10Mbps.  Even going into Device Manager and setting it to 100Mbps, or 1000Mbps, Half or Full duplex does nothing.  It disconnects the Link and Reconnects but only sits at 10Mbps.  Im normally connected via the IntelAMT via RealVNC.

Ive updated from the 0054 bios to 0055 last week so Im also on the latest Bios revision.

What causes this to get in such a stuck state that only killing power to the board fixes this?

Dixit

My network is set up with 8021q tagging enabled, and this works perfectly well while in an operating system, but I can't seem to vlan tag during PXE boot

Is there any way, using bootutil or the like, to set the native vlan during PXE boot on Intel cards? (Specifically an Intel 82599EB 10gb card, but I'm assuming it will be a generic way)

Thanks,

Sean

My Intel motherboard has an PRO1000PL lan-on-motherboard with 82573L controller (device id 109A).

The latest Bios includes Intel AMT Bx v1.2.3. However I am not getting the 'Ctrl - P' prompt to enter iAMT setup.

I believe this is due to the specification for the 82573L, which is not AMT enabled/capable. The 82573E however is AMT capable.

Is this correct?

Is it a software/firmware limitation or a hardware limitation?

Also, I have the option in Bios, under 'Management' section, to enable or disable ASF (ASF v2 according to datasheets). However I havent been able to find any configuration documentation for ASF.

Is there anything about configuring ASF2? Technical reference? Configuration Guide?

Thanks.

Hello,

I recently installed Windows Server 2008 R2 for testing purposes but it's missing the Ethernet Controller driver and the driver that I've downloaded from Intel's database will not detect the card.

I had the same problem with Windows Server 2012 previously and I solved it by editing the e1c63x64.inf

I found the instructions here

I tried the same thing on Win Server 2008 but it did not work. I'm wondering if I have to edit a different file or something and if so, which file would that be?

Thanks.

Hi There,

Our NICs are running driver v9.13.41.0 and by default it appears the driver does not strip VLAN ID/802.1Q headers, however our Sniffer (Observer) does not pick up the VLAN ID.

I have followed the instructions set out here and modified the registry, but we still cannot see the VLAN information. I'm wondering if this problem is related to 2008 x64 and the driver v9.13.41.0

Has anyone encountered the same issue or is there a setting I am missing ?

Server Information:

Observer Version: 14 Build: 0005.0000

OS: Windows 2008 R2 Standard – 64BIT

NICs:

Intel(R) PRO/1000 PF Server Adapter - 9.13.41.0

Intel(R) PRO/1000 PT Dual Port Server Adapter – Driver Version 9.13.41.0

All registry changes were made to the following location:

HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4D36E972-E325-11CE-BFC1-08002BE10318}

Reference guides followed:

http://www.intel.com/support/network/sb/CS-005897.htm

Hi guys,

I am having a problem with the igb driver on FreeBSD

Issue:

We have two Dell R720 servers with the Intel i350 QP adapter.

We are need three connections.

1. Connection to external switch

2. Connection to internal switch

3. Connections to the other server via crossover cable

Connection 1 and 2 are working fine and remain consistent and are connected to a Cisco switch.

But connection three keeps dropping out and when we try bring it back online it doesn't respond.

We have restarted the networking on both servers and the connection will come back for about 10 minutes then drop again. This is also the case if we restart the servers.

We have tried a straight through cable instead of the crossover with the same effect.

I have contacted Dell and they said they cannot help because Intel develop the driver for FreeBSD and i should contact them, but when i contact Intel they tell me they only support Intel boxed products.

Anyone got any ideas?

I have 3 PC's, each of which is running Windows 7 x64, and each of which had no prior problems connecting to the internet or the network via their PCI-e Intel Gigabit CT Desktop Adapters.  Over the course of the last week, each PC started connecting to "Network 2", instead of the usual Network, and while they could connect to the internet, they would not see the other PCs on the network or the Windows Home Server.  I tried disabling the adapter, deleting the Network and Network 2 profiles, and then re-enabling the adapters.  While this allowed the adapters to connect to the Network Profile, none of the other PCs on the network or the WHS could be accessed, even though I still had full internet connectivity.

When I disconnected the ethernet cables from the Intel CT adapters, and connected them to the embedded adapters on each PC's motherboard, the network was instantly discovered, so I thinking the culprit must be the CT adapters.  However, it can't be possible that all three failed at the same time, or in the same manner.  There has to be something else going on...

I'm certainly no networking expert, so I would appreciate any help from the community.  I paid a decent premium for these adapters, and all are using the most recent drivers.  Suggestions?

Thanks,

Demaximis

<><> Apologies if this has been posted in the wrong place <><>

Ok I'll keep this brief.

Problem:- LAN regulary disconnects and reconnets 30 seconds later.

Symptons:- Loose connection to Lan / Internet for around 30 seconds.

Background:- Had this problem with both P8P67 B2 and P8P67 pro boards. I have several other computers connected to the switch (not hub) and they are working fine. Have replaced the lead to no avail. Have even used the same lead in several other computers which works fine.

Config:-

Study

Netgear 8 port SWITCH 10/100/1000
Server
VOIP
PC

Switch connected to lounge hard wired via outdoor sheilded cat6 lead.

Lounge

Netgear 8 port SWITCH 10/100/1000
Router
XBOX 360
Wii
<><><>

Message in System Event Logs;-

<><>
Warning message - date time - source = e1cexpress
Event ID = 27

Intel 82579V Gigabit Network Connection
- Network link is disconnected
<><>

Then it states its connected again.

Have also tried the following;-

Remove Kaspersky 2011
Ensure ALL power management even OS is disabled
Use IPV4 instead of IPv6 in prefix policies
Disable nativue IPv6
Disable tunnel IPv6
Disable IPv6

netsh interface tcp set global rss=disabled
netsh interface tcp set global autotuninglevel=disabled
netsh int ip set global taskoffload=disabled

Disabled SNP;-
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Servic es\Tcpip\Parameters]
EnableTCPChimney=dword:00000000
EnableTCPA=dword:00000000
EnableRSS=dword:00000000

Have tried driver from Asus MB CD, Asus Website, Your Website, Windows Update all to no avail.

Please help.

Hi,

For a XEN environment I have doubts if SR-IOV will provide the required features compared to nPAR.

Servers will be Dell M620 with either Intel X520 or Broadcom 57711, running Linux kernel 3.8, XEN 4.2 and openvswitch 1.7

I need separate networks on the hypervisor:

- Management

- Storage

- Openvswitch

- Passthrough for VMs

So my questions are:

- Can I use the PF's on the hypervisor?

- Can I use 4 VFs on the hypervisor and the rest for VM passthrough?

- Can I bond VFs with LACP?

- Can I use a VF as a iSCSI HBA?

- Is DCB supported in above configuration?

- Can I guarantee 1 VF 50% and let the rest use the default round robin fairness?

Hope someone is able to help me with answers or documentation, as I have had difficulties finding it.

Thanks.

Regards

Kristoffer

I'm using an OS reserverd memory as DMA mapping memory, and then map it to Guest OS. at the same time, the VT-d is enabled.

As the VT-D spec indicated, I set up an identiry mapping for reserverd memory , it works well in PF, but after I change to VF, the following error log are outputted :

[   95.030756] DRHD: handling fault status reg 2

[   95.030764] DMAR:[DMA Write] Request device [03:10.0] fault addr 0

[   95.030766] DMAR:[fault reason 05] PTE Write access is not set

[   95.030772] DMAR:[DMA Write] Request device [03:10.0] fault addr 0

[   95.030773] DMAR:[fault reason 05] PTE Write access is not set

[   95.030779] DMAR:[DMA Write] Request device [03:10.0] fault addr 0

[   95.030780] DMAR:[fault reason 05] PTE Write access is not set

[   95.030787] DMAR:[DMA Write] Request device [03:10.0] fault addr 0

[   95.030788] DMAR:[fault reason 05] PTE Write access is not set

Did anybody encount this kind of issue?

my enviroment:

Kernel: 3.0.4

Guest OS: kernel 2.6.34
CPU : Intel xeon
Hypervisor : qemu 1.0.1

All,

I was wondering if there is a timed mechanism built into the embedded 4-port I350 that would affect the ISCSI boot process.

I currently have an issue where the I can successfuly boot from say NIC Port0, but if I move my connection to NIC Port3 and configure the attempt again, it will not establish a connection.  It cannot login and the IP cannot be hit with a ping.

Here is the setup: 

IBM 3650 M4 w/ UEFI configured with 1 attempt for NIC3. IQN settings as well as IP's verified for the Host and SAN.

Netapp 3170 w/ a LUN containing a Windows 2008 R2 OS, mapped and presented to the IBM

Both connected to a Brocade MLXe Switch.

Let me know if you need more info.

Thanks in advance!

We are configuring VFs for our high-volume network traffic, but want to continue to use Linux bridging for other VMs. Is there some configuration we need to do to enable packets from a VF to be switched to a bridge off the eth0 PF?

We create several VMs. One of them has a high volume of traffic, as it serves as a load balancer. It takes external traffic on one VLAN, and distributes it to other VMs on a different VM. If the different VM is on the same host, the packets are never switched in the 82599 back to the eth0 for feeding into the Linux bridge. We are using Red Hat EL 6.4. We have been able to get VM to VM traffic working on the same host, if both VMs use VFs, but are unable to get VM to VM traffic if one is using VF and the other is using bridges.

Suggestions?

I have 3 client PCs running various versions of Windows and 1 file server, running Vista Home Premium 64.  All 4 machines are using the Gigabit CT Desktop Adapter for connectivity.  This server has previously been running Linux for 2 years and no hardware has changed, but I just installed Windows instead.  My problem is this:  WOL from any of the clients works perfectly, as far as waking up the server.  However, I am then unable to connect to the server.  When I go to the server and look in Device Manager, the network adapter has the exclamation point/triangle thing.  If I disable the adapter in Device Manager, and then enable it, everything works perfectly. This happens every time it wakes from sleep mode, whether using WOL or waking it directly from the keyboard.

I never had this issue while running Linux, so naturally I suspect a Windows configuration issue or driver issue.  Except, one of the 3 client machines is an identical copy, hardware and software (yes, and drivers), for the server, and it never has a problem when waking up.

This behavior started immediately after installing Vista so I really don't believe it's a hardware issue, although I suppose this weekend I could swap card to 100% rule it out.  But if anyone has seen this behavior before, and/or has any ideas or suggestions on what I should look for to fix it, I'd appreciate it.

Thanks!

Is it possible to set up adapter teaming / bonding on two VFs from two Intel X520-T2 ports? I would like to add resiliency to my VMs....

I have two VFs within my virtual Windows Server 2008 R2 under Citrix XenServer 6.0 and simple networking works all fine for both, but the usual adapter teaming options within the Intel driver settings are not visible on the VF adapter properties.

Any Help?

Thanks in advance

Marco

Hi, I'm stuck with hardware that has the 82598 10G controller instead of the 82599. I know that it doesn't support hardware filtering, but my requirements are a little simpler than that: I need to put specific destination IPs or MAC addresses into their own receive queue so I can read from them on different cores. It looks like I should be able to do this by either using RSS or VMDq, but the datasheet is fairly vague on how to set things up. I've scoured the internet trying to find how to do this, but it seems that nobody talks about the 82598. PF_RING and DPDK don't seem to support this, so any examples or advice would be appreciated. Thanks!

Hi All,

   I am new to this community and this is my very first post. Hope I am asking right question in the right place. If not please let me know where to post my query.

    I am characterizing one Intel(R) Server Board S2600IP, for network processing mainly handling packets.

The board has following configuration:

    a. Intel® Xeon® Processor E5-2620 (15M Cache, 2.00 GHz, 7.20 GT/s Intel® QPI)

    b. 12 physical cores on 2 sockets (6 cores on each socket)

    c. an Ethernet Server Adapter I350-T2

    d. Installed Linux (kernel - 2.6.35.14)

I am using the latest Intel igb driver, version 4.1.2 and using iperf for packet injection but observing that the bandwidth performance is not upto the mark.

The driver parameters:

    a. IntMode=2,2

    b. RSS=6,6

smp affinity:

        Core Id 6 : eth(0,1)-RxTx-0

        Core Id 7 : eth(0,1)-RxTx-1

        Core Id 8 : eth(0,1)-RxTx-2

        Core Id 9 : eth(0,1)-RxTx-3

        Core Id 10 : eth(0,1)-RxTx-4

        Core Id 11 : eth(0,1)-RxTx-5

But I am getting very poor throughput not more that 121 Mbps when the device is in bridge mode.

Moreover the cores are not (almost) evenly loaded. Only one core is mostly busy 100% (si value) others are maximum 2% used.

Please help me or suggest me the parameters, both for OS and the driver, to get the best performance.

Waiting for your early responses.

Best regards,

--Chinmay

New Windows 8 install on DZ77GA-70K board.  About every 24-48 hours the 82574L integrated NIC fails hard--all communication through it stops and Windows will not shut down--the NIC seems to be hanging it.  I see the same thing when looking at the NIC properties when it is hung and trying to apply any changes (e.g. IP address change)--the properties dialog hangs.

Have updated to latest drivers and latest BIOS.  Problems continue.  Am about to RMA the board.

Hi, I'm stuck with hardware that has the 82598 10G controller instead of the 82599. I know that it doesn't support hardware filtering, but my requirements are a little simpler than that: I need to put specific destination IPs or MAC addresses into their own receive queue so I can read from them on different cores. It looks like I should be able to do this by either using RSS or VMDq, but the datasheet is fairly vague on how to set things up. I've scoured the internet trying to find how to do this, but it seems that nobody talks about the 82598. PF_RING and DPDK don't seem to support this, so any examples or advice would be appreciated. Thanks!

Hi Folks,

We are trying to bond two SRIOV VF's (coming out of two different PF's) within a Linux VM in active loadbalance mode using linux bonding driver. This is primarily to have high-availability across the links/switches & without losing out on the bandwidth/performance of the links.

One of the issue we hit upon is the driver constantly printing "Spoofed packets detected" & the bond doesnt work at all. Is this is due to the active-load-balance mode sending the same mac across the VF's & at HW level they are considered as spoofed packets? Is recompiling ixgbe with mac-spoofing disabled the only option for this? Or is there some way to get around this problem?

We in fact tried to change to active-backup mode with fail_over_mac (in bonding driver) set to active/follow modes. But even in that case the failovers dont seem to happen.

Have anybody tried bonding SRIOV VF's within Linux VM (especially active-load-blance) & got it working? If so, can you pls share your thoughts on this?

Thanks.

--Shyam