Network Problem with newest generation of lenovo M920q

[RLadi]

We have 16 Lenovo M910q running with Ubuntu 16.04 without any Problems. Now we bought 20 M920q and are running into problems with network connection. After 16.04 didn't work we installed 18.04.

A ping to an address in our network showed some 1000 ms answer time.

After installing the hwe-kernel with linux 4.18 and compiling the network driver things looked a little bit better.

For some minutes connectivity was ok, answer times were under 1ms, but some time later answer times changed to some 100ms and 1000ms.

Is there a known problem with the intel I219LM (rev 10) in connection to linux?

[Crisselle_Intel]

Hello RLadi, Thank you for posting in Intel Ethernet Communities. Please provide the System Support Utility log from Lenovo M910q and M920q. This will allow us to check your Adapter details and configuration. Please refer to the steps below: 1- Download from https://downloadcenter.intel.com/download/26735/Intel-System-Support-Utility-for-the-Linux-Operating-System 2- Open SSU.exe 3- Mark the box "Everything" and then click "Scan". 4- When finished scanning, click "Next". 5- Click on "Save" and attach the file to a post. Let us know if you have any other questions. Best regards, Crisselle C Intel Customer Support Under Contract to Intel Corporation

[RLadi]

Thank you for your answer,

attached are two files for M910q (Ubuntu 16.04) and M920q (with 18.04 and HWE-kernel 4.18 )

[RLadi]

Thank you for your answer,

attached are two files for M910q (Ubuntu 16.04) and M920q (with 18.04 and HWE-kernel 4.18 )

[Crisselle_Intel]

Hello RLadi, Thank you for the information. We will check on this and we will get back to you as soon as possible. Let us know if you have any other questions. Best regards, Crisselle C Intel Customer Support Under Contract to Intel Corporation

[Crisselle_Intel]

Hello RLadi, Thank you for waiting for our update. May we know if you already tried the latest driver from sourceforge(Please note this is third party website, we have no control over the content therein) https://sourceforge.net/projects/e1000/files/e1000e%20stable/ Let us know if you have any other questions. Best regards, Crisselle C Intel Customer Support Under Contract to Intel Corporation

[RLadi]

Hello CrisselleC, I used driver 4.3.2.1 from intel website; I will check driver from sourceforge and report. RLadi Zitat von Intel Forums <supportreplies@intel.com>:

[Crisselle_Intel]

Hello RLadi, Thank you for the reply. We will wait for your update. Let us know if you have any other questions. Best regards, Crisselle C Intel Customer Support Under Contract to Intel Corporation

[RLadi]

Hello, I compiled the driver from sourceforge (4.3.2.3 ) and saw the same problem: example: admin@O2-10-PCT5:~$ ping boss PING boss.borg-ibk.local (192.168.200.200) 56(84) bytes of data. 64 bytes from boss.borg-ibk.local (192.168.200.200): icmp_seq=1 ttl=63 time=1008 ms 64 bytes from boss.borg-ibk.local (192.168.200.200): icmp_seq=2 ttl=63 time=3064 ms 64 bytes from boss.borg-ibk.local (192.168.200.200): icmp_seq=3 ttl=63 time=2632 ms 64 bytes from boss.borg-ibk.local (192.168.200.200): icmp_seq=4 ttl=63 time=3149 ms 64 bytes from boss.borg-ibk.local (192.168.200.200): icmp_seq=5 ttl=63 time=2667 ms 64 bytes from boss.borg-ibk.local (192.168.200.200): icmp_seq=6 ttl=63 time=3377 ms 64 bytes from boss.borg-ibk.local (192.168.200.200): icmp_seq=7 ttl=63 time=2664 ms 64 bytes from boss.borg-ibk.local (192.168.200.200): icmp_seq=8 ttl=63 time=3677 ms 64 bytes from boss.borg-ibk.local (192.168.200.200): icmp_seq=9 ttl=63 time=4552 ms 64 bytes from boss.borg-ibk.local (192.168.200.200): icmp_seq=10 ttl=63 time=4722 ms 64 bytes from boss.borg-ibk.local (192.168.200.200): icmp_seq=11 ttl=63 time=3706 ms 64 bytes from boss.borg-ibk.local (192.168.200.200): icmp_seq=12 ttl=63 time=2706 ms 64 bytes from boss.borg-ibk.local (192.168.200.200): icmp_seq=13 ttl=63 time=1690 ms 64 bytes from boss.borg-ibk.local (192.168.200.200): icmp_seq=14 ttl=63 time=690 ms 64 bytes from boss.borg-ibk.local (192.168.200.200): icmp_seq=15 ttl=63 time=521 ms 64 bytes from boss.borg-ibk.local (192.168.200.200): icmp_seq=16 ttl=63 time=1001 ms 64 bytes from boss.borg-ibk.local (192.168.200.200): icmp_seq=17 ttl=63 time=520 ms 64 bytes from boss.borg-ibk.local (192.168.200.200): icmp_seq=18 ttl=63 time=249 ms But then I found, that everything is ok ( answer times below 1 ms) when I connect to another switch. * Network connection is o.k. with 100Mbit Switch Cisco 2960 * Problem occurs with Gbit Switch Cisco XG350-28 To me the main difference is autonegotiation or mdi-x auto? thanks for your help RLadi Zitat von Intel Forums <supportreplies@intel.com>:

[Crisselle_Intel]

Hello RLadi, Thank you for the update. What is the firmware version installed on your Cisco switch? Are you using the latest firmware? What is the speed mode configured on your switch? Since you mentioned that the issue you've encountered occurred on your switch, may we know if you tried to check this with Cisco? There might be some issues that's the same on what you've encountered. Let us know if you have any other questions. Best regards, Crisselle C Intel Customer Support Under Contract to Intel Corporation

[RLadi]

Hello CrisselleC, hello Madsen, Now I updated the firmware of the Cisco SG-350-28 to the newest image and everything is ok. I did a new install of Ubuntu 18.04 with the default kernel (4.15.0-45-generic) and everything worked. It was not necessary to install a newer kernel or a newer network-module ( e1000e) . When I connect the M820q to another Switch - Cisco C2960S, i still have the same problem. The Firmware on this Cisco C2960 is about four years old. The next days I will try to update the firmware of this C2969S and do some testing. But the whole problem only arises with Intel I219-LM (rev 10) [ 8086:15bb ] . With I219-LM [ 8086:15b7 ] there is no problem with my Gbit-Switches even with old firmware. Best regards RLadi Zitat von Intel Forums <supportreplies@intel.com>:

[madsen]

I seem to have the same problem.

When freshly booted the connection is fast and ping times are sub ms. After a while speed drops to even below 20kbps (not kBps) and ping times might go skyhigh (they don't always do that). At one time disabling-enabling the NIC seemed to help. WiFi works as it should.

Tried it with a non managed 1Gbps switch but had the same problem, disconnecting-reconnecting the cable helped once with that switch.

Normally connected to a cisco c2960s (configured auto and has negotiated to 1000Mb/s). There are no problems with other computers connected to that switch and the port I'm connected to was working fine with the m900 that I just replaced.

I've attached the ssu scan file.

[Crisselle_Intel]

Hello RLadi, Thank you for the update. We will wait for the the results after your Cisco C2960 firmware has been updated. @madsen May we know if you tried the latest driver from sourceforge? Let us know if you have any other questions. Best regards, Crisselle C Intel Customer Support Under Contract to Intel Corporation

[madsen]

The latest driver (3.4.2.3 - 29.1.2019) seems to have helped.

I've got 15.2(2)E4 on my cisco switches and upgrading those would be a PITA since we have stacks with different models so finding the common software version is usually really hard.

I'll be back posting in case the problems continue.

[RLadi]

Hello Crisselle, hello madsen, 1. I updates my Cisco Stack 2x2960S to firmware 15.2(2)E9; The M920q with Standard Ubuntu 18.04 showed with the C2960 the same problems. Next I set the connection speed on the from auto (1000) to 100 Mbit. Now everything worked. 2. Next I updated 18.04 to the hwe-kernel (kernel 4.18.0-15-generic) . The problem remains. Working with administratively set 100Mbit, not working with 1000Mbit (auto). 3. Now I installed the newest driver (e1000e-3.4.2.3). root@inf-1-02:~# modinfo e1000e filename: /lib/modules/4.18.0-15-generic/updates/drivers/net/ethernet/intel/e1000e/e1000e.ko version: 3.4.2.3-NAPI Result: The same as above: Not working with auto (1000). Not working with 1000 (administratively set). Working with 100Mbit (administratively set) At the moment I can work with my M920q connected with 100Mbit and I will wait for updated drivers. Best regards, RLadi Zitat von Intel Forums <supportreplies@intel.com>:

[Crisselle_Intel]

Hello RLadi, Thank you for the update. May we know if you tried contacting Lenovo about this issue and if they have driver updates? Let us know if you have any other questions. Best regards, Crisselle C Intel Customer Support Under Contract to Intel Corporation

[Crisselle_Intel]

Hello RLadi, Please let us know if you were able to contact Lenovo to check if they have newer driver version that could be of help? If you have any other questions, please do not hesitate to ask. Best regards, Crisselle C Intel Customer Support Under Contract to Intel Corporation