[SOLVED] Proxmox VE Installer doesn't show eno1, any ideas?

hohl

Member
Jun 18, 2020
33
2
13
30
I just wanted to install Proxmox VE 6 on a Dell PowerEdge R630 with 4 built-in network ports (Intel X520/I350: 2x SFP+ and 2x RJ45). I only connected the first SFP+ which iDRAC shows me as port 1 and as the only active port. However, when I launch the Proxmox VE Installer from the ISO it only shows me all the inactive network interfaces besides eno1:

Screenshot 2020-08-04 at 11.51.46.png

Any ideas what I am doing wrong or how I can fix that?

(The MAC address of the connected network link is the one ending in :98 according to iDRAC which isn't shown, so I am certain that I'm missing the eno1 here.)
 
Last edited:
I just wanted to install Proxmox VE 6 on a Dell PowerEdge R630 with 4 built-in network ports (Intel X520/I350: 2x SFP+ and 2x RJ45). I only connected the first SFP+ which iDRAC shows me as port 1 and as the only active port. However, when I launch the Proxmox VE Installer from the ISO it only shows me all the inactive network interfaces besides eno1:

View attachment 18981

Any ideas what I am doing wrong or how I can fix that?

(The MAC address of the connected network link is the one ending in :98 according to iDRAC which isn't shown, so I am certain that I'm missing the eno1 here.)
Is your Dell server using the first NIC port as a dedicated iDRAC port perhaps?
 
could you try starting the installer in debug mode and check the output of `ip link show` in the second debug shell?
does eno1 show up there?
 
Thanks for the suggestion, but no, it doesn't show there either; again only eno2 to eno4 (and lo).
 
in that case check the output of `dmesg` in that debug shell (you should be able to scroll upwards with 'shift+pgup') check for messages from the network drivers (especially ixgbe)

on a hunch I would guess that this is related to an unsupported sfp being used in the NIC

I hope this helps!
 
as said- see the line - 'unsupported SFP+ or QSFP module type was detected'

as a possible workaround you could try to:
Code:
rmmod ixgbe
modprobe ixgbe allow_unsupported_sfp=1
dmesg

in the second debug shell - if the nic then gets recognized you can install with it and redo those steps in the installed system after installation+reboot
(It would probably be more robust to find a supported SFP+ module, since the paramter-type can change between kernels and it does cause inconveniences if your link stays down after an upgrade)

I hope this helps!
 
Thank you very much, especially since the error wasn't really related to Proxmox, but me just being unaware how picky Intels SFP+ adapters seem to be. That flag indeed made it work, but I will anyhow try to replace the SFP+ module with one which should be officially supported. (Just so I can sleep well, as they don't seem overly expensive anyhow.)
 
Well, more or less: the installer was happy with the interface and the IP configured. After rebooting I created a file "/etc/modprobe.d/ixgbe.conf" with options ixgbe allow_unsupported_sfp=1 in it and then run sudo update-initramfs to persist it for the future. ip link show also shows it as "UP" and "DEFAULT", but not even a simple ping 8.8.8.8 is successful. The IP/Netmask/Gateway should be right - checked that already multiple times - dmesg also doesn't show any errors related to the network.

Does the installer even make use of the network? Or might it just have seemed like the network worked (because there weren't any errors at all), but it actually just loaded the network interface without ever working/getting a connection?
 
Does the installer even make use of the network?
yes for example to determine the location of the server (e.g. for debian mirror selection)

ip link show also shows it as "UP" and "DEFAULT",
this just states that the interface is set to up by the admin - not that an actual link is detected (look for LOWER_UP)

please post the output of:
Code:
ip address show
ip route show
cat /etc/network/interfaces

and check dmesg again
 
yes for example to determine the location of the server (e.g. for debian mirror selection)
Good to know, because then at least it should work from a hardware side.

please post the output of:
Here we go:

ip address show
Screenshot 2020-08-05 at 09.41.44.png

ip route show
Screenshot 2020-08-05 at 09.43.54.png

cat /etc/network/interface
Screenshot 2020-08-05 at 09.45.11.png

The only related lines found through dmesg:
Screenshot 2020-08-05 at 09.49.25.png

All red blocks above hide the same IPv4 prefix. And, yes - the interface-config was already altered by me to use routing for the VMs interfaces, but it didn't work before with the default config (the bridged one) neither. IP/Netmask/Gateway used for the Proxmox management interface should be right, because I use the same subnet with just a slightly different IP for the iDRAC interface and that one is online and usable.
 
can you ping your gateway?
ping xx.xx.77.1?
if not check the arp-table (ip neigh show)

else everything looks ok from a quick glance
 
No...

Screenshot 2020-08-05 at 10.08.16.png

(I also tried pinging X.Y.77.87 too, so don't be surprised about that one in the next)

Screenshot 2020-08-05 at 10.13.24.png
 
* I guess the issue is somewhere in your network - check the cables, switches etc.
* check on your gateway if the icmp/arp requests arrive there
* check with tcpdump on the PVE-node if you see any traffic coming in
 
* I guess the issue is somewhere in your network - check the cables, switches etc.
* check on your gateway if the icmp/arp requests arrive there
The network is provided as part of a colocation agreement and I don't have access to their switches, so that's gonna a bit hard. I actually don't assume the error is on their site, but I guess I will have to contact anyhow.

* check with tcpdump on the PVE-node if you see any traffic coming in
Nothing besides the outgoing ARP requests:

Screenshot 2020-08-05 at 10.31.36.png
 
The network is provided as part of a colocation agreement and I don't have access to their switches, so that's gonna a bit hard. I actually don't assume the error is on their site, but I guess I will have to contact anyhow.
would do so in any case - also the STP packets would indicate that some kind of link is present - maybe the port is just in the wrong VLAN....
 
would do so in any case

It was indeed a miss-configuration on their end (and they also fixed the above issue of the complaining "unsupported module" for my picky Intel X520). So, thank you one more time for helping me with an issue which wasn't even related to Proxmox itself and sorry for wasting your time. Anyhow, best regards from Graz and thanks to your team for providing such an amazing software.
 
  • Like
Reactions: Stoiko Ivanov
Glad the issue was found - was not a waste of time!
Thanks for the compliment - I'll relay it around the office - greetings from Vienna :)

Please mark the thread as 'SOLVED' - that can help others running into a similar issue
 

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!