Network card TPLINK doesn't work on 2 nodes

frankz

Active Member
Nov 16, 2020
364
24
38
CIao at all, I bought 3 tplink network cards model TG 3468 and tried to install on cluster nodes.

On a knot it works and on 2 no. It happens that after inserting the card then express on the only one available or the 16x on both computers the cards currently in production cease to work, that is, the links do not turn on. This happens before the operating system is loaded and after, and as if they were going into conflict as soon as the card is inserted on the bus then express 16 x. Only one of the three despite having inserted the card on the 16x available bus, everything works.

The models for which I have the defect is a Fujitsu TX 1310 M3 4 x Intel(R) Xeon(R) CPU E3-1225 v6 @ 3.30GHz (1 Socket) and the other an DELL t40 4 x Intel(R) Xeon(R) E-2224G CPU @ 3.50GHz (1 Socket) . In addition, in the system logs I did not find anything in particular as the cards are recognized but I exclude those already configured.
 

Attachments

  • bus2.jpg
    bus2.jpg
    426.7 KB · Views: 21
It is possible, that the enumeration of the already configured NICs has changed by placing an additional PCI card. Check the output of ip a against your /etc/network/interfaces file.
 
It is possible, that the enumeration of the already configured NICs has changed by placing an additional PCI card. Check the output of ip a against your /etc/network/interfaces file.
Yes Aaron, I can do it but the thing that doesn't convince me much and that before the boot phase I see on the switch the link of the previously mounted cards turn on to 10 mb and then turn off after the system starts starting to start off! However, looking at the status of the network on the node I only see that the new interface is enumerated with enp6s0 the others remain active but not working, in particular the one dedicated to cluster data transfer. Thank you for the moment. Ps: a similar situation happened to me with an intel gigabit pCI that I had to return as it gave a similar defect .....
 
Yes Aaron, I can do it but the thing that doesn't convince me much and that before the boot phase I see on the switch the link of the previously mounted cards turn on to 10 mb and then turn off after the system starts starting to start off! However, looking at the status of the network on the node I only see that the new interface is enumerated with enp6s0 the others remain active but not working, in particular the one dedicated to cluster data transfer. Thank you for the moment. Ps: a similar situation happened to me with an intel gigabit pCI that I had to return as it gave a similar defect .....
Hi, I changed the network card to a PCI but you don't want to know how it works. Do you have any idea what I can control? I noticed that adding the vmbr3 in the Interfaces file missed hwaddress that I added.
Anyway, I'm trying them all but it seems that the two servers over 3 network adapters don't want to take ..... hi
I made other changes and managed to make the Link active:

Ethtool-s enp6s0 speed 1000 duplex full autoneg off
I don't know exactly how to insert this configuration in order to restart the node .
However, it still doesn't work because the kernel crashes ..
 

Attachments

  • Schermata 2021-12-31 alle 12.43.57.png
    Schermata 2021-12-31 alle 12.43.57.png
    64.4 KB · Views: 8
  • Schermata 2021-12-31 alle 12.43.38.png
    Schermata 2021-12-31 alle 12.43.38.png
    131.1 KB · Views: 8
  • Schermata 2021-12-31 alle 12.47.57.png
    Schermata 2021-12-31 alle 12.47.57.png
    219 KB · Views: 8
  • Schermata 2021-12-31 alle 19.15.27.png
    Schermata 2021-12-31 alle 19.15.27.png
    61.1 KB · Views: 6
  • Schermata 2021-12-31 alle 19.30.05.png
    Schermata 2021-12-31 alle 19.30.05.png
    294.1 KB · Views: 6
Last edited:
Hi, I changed the network card to a PCI but you don't want to know how it works. Do you have any idea what I can control? I noticed that adding the vmbr3 in the Interfaces file missed hwaddress that I added.
Anyway, I'm trying them all but it seems that the two servers over 3 network adapters don't want to take ..... hi
I made other changes and managed to make the Link active:

Ethtool-s enp6s0 speed 1000 duplex full autoneg off
I don't know exactly how to insert this configuration in order to restart the node .
However, it still doesn't work because the kernel crashes ..
Given that the r8169 chip doesn't want to know how to work in some contexts, what does anyone think if I buy Intel 82575eb?
 
However, for those interested in the problem, I solved with two network cards with INTEL 82576 chip. I had to reassociate the name of the Neth in the bridge configurations, but everything works at the moment. Therefore and personally I can say that in some contexts there may be malfunction problems with REALTEK chip cards!
 

About

The Proxmox community has been around for many years and offers help and support for Proxmox VE, Proxmox Backup Server, and Proxmox Mail Gateway.
We think our community is one of the best thanks to people like you!

Get your subscription!

The Proxmox team works very hard to make sure you are running the best software and getting stable updates and security enhancements, as well as quick enterprise support. Tens of thousands of happy customers have a Proxmox subscription. Get yours easily in our online shop.

Buy now!