x710-DA2 iSCSI Connections Keep Disconnecting

[BTech3]

Our setup:

• Server 2012 R2 Hyper-V Cluster
• 3 HP DL380p Hosts
• HP MSA 2040 SAN
• 2 Cisco Nexus 5000 series Switches dedicated to iSCSI storage network
• Each Host has an Intel x710-DA2 10Gb NIC, latest firmware and drivers
• Each Host has 1 10Gb iSCSI over fiber connection to each Switch
• Switches are not Stacked or connected
• 2 iSCSI subnets, one for each Switch
• SAN has 4 10Gb connections, 2 for each controller, one iSCSI subnet on each controller
• Host iSCSI Initiator set for MPIO, each Host IP connects to both SAN IPs
• MPIO set to Round Robin with Subset

We are getting these errors in the Windows Event Log on each Host:

• Event ID 20: Connection to the target was lost. The initiator will attempt to retry the connection.
• Event ID 7: The initiator could not send an iSCSI PDU. Error status is given in the dump data.
• Event ID 34: A connection to the target was lost, but Initiator successfully reconnected to the target. Dump data contains the target name.
• Event ID 27: Intel(R) Ethernet Converged Network Adapter X710 Network link is disconnected.
• Event ID 31: Intel(R) Ethernet Converged Network Adapter X710 Network link has been established at 10Gbps full duplex.
• Event ID 27: Intel(R) Ethernet Converged Network Adapter X710-2 Network link is disconnected.
• Event ID 31: Intel(R) Ethernet Converged Network Adapter X710-2 Network link has been established at 10Gbps full duplex.

[SYeo3]

Hi BCT-Tech,

Thank you for contacting Intel.

I understand you are using fiber connection on your X710 NIC. Can you provide the brand/model of your fiber module (SFP+)?

In the meantime,please refer to the iSCSI Quick-Connect Guide below:

http://www.intel.com/content/dam/www/public/us/en/documents/guides/iscsi-quick-connect-windows-guide.pdf http://www.intel.com/content/dam/www/public/us/en/documents/guides/iscsi-quick-connect-windows-guide.pdf

Please download run the Intel® System Support Utility (SSU) so we may see your NIC's configuration.

Websites:

1. http://www.intel.com/support/software/applications/ssu/sb/CS-035779.htm Intel® System Support Utility (Intel® SSU) Frequently Asked Questions

2. https://downloadcenter.intel.com/download/25293/Intel-System-Support-Utility Download Intel® System Support Utility (SSU)

Click on SCAN to scan your system information and SAVE to save it to a file. You may attach the scan results here.

Hope this helps.

Sincerely,

Sandy

[BTech3]

Sandy_Intel wrote:

Please download run the Intel® System Support Utility (SSU) so we may see your NIC's configuration.

Websites:

1. http://www.intel.com/support/software/applications/ssu/sb/CS-035779.htm Intel® System Support Utility (Intel® SSU) Frequently Asked Questions

2. https://downloadcenter.intel.com/download/25293/Intel-System-Support-Utility Download Intel® System Support Utility (SSU)

Click on SCAN to scan your system information and SAVE to save it to a file. You may attach the scan results here.

Because of the disconnects, we have reverted back to using the 1Gb Cat5e connections to the SAN. The 10Gb fiber connections are still enabled, just taken out of the iSCSI Initiator until this issue is resolved. The SFP transceivers are Intel, but I'm not sure of the exact model. I will try and get that for you (should be E10GSFPSR). Attached is the SSU results from one of the Hosts.

We just recently noticed that the disconnects also occur when we open up the NIC properties and press the Configure button to view the NIC settings (picture attached).

[BTech3]

We have also tried installing just the base drivers, and also the Proset / Advanced Network Services features. The disconnects happen with both.

Attached is the SSU results from another Host.

[BTech3]

Anyone?

[st4]

Hi BCT-Tech,

Thank you fo the info. We will look into this.

rgds

wb

[SYeo3]

Hi BCT-Tech,

The event IDs show the link is disconnected while we are unable to investigate since the SSU doesn't show the settings of the NIC that is in question. As of the moment, please try to adjust the advance settings to see if it can help improve the connection. See details below:

1. Decrease Interrupt Moderation Rate

2. Increase Transmit and Receive Descriptors - Increase the Transmit Descriptors buffer size. The default is 256. Maximum value is 2048.

3. Disable TCP/UDP Checksum Offload

Kindly refer to the website below for the list of Adapter's advance settings:

http://www.intel.com/support/network/adapter/pro100/sb/CS-029402.htm Intel® Server Adapters — Advanced Driver Settings for Intel® Ethernet 10 Gigabit Server Adapters

In case you continue to encounter the same issue, please contact HP to check for the list of compatible hardware of your server model.

Sincerely,

Sandy

[JJonC]

We are seeing similar disconnects on x710 adapters in Dell servers. Steps 1 & 2 did not resolve the issue, but disabling offloading did. We would prefer to utilize the offloading functionality of these NICs. Is there another solution other than disabling offloading?

Thanks,

Jon

[tvuor]

Hi, did you figure out the cause of this problem? I seem to have the same, with our X710 nics.