Unable to determine IP address of primary interface

7 posts / 0 new
Last post
#1 Mon, 09/17/2018 - 23:20
MichaelS

Unable to determine IP address of primary interface

Hi Everyone.

Been using Virtualmin for some years for various projects, and basically have zero issues with installations...until now.

The install script errors near the end of Step 3 not being able to find the IP. Here is the error below;

Device "link" does not exist. [WARNING] Unable to determine IP address of primary interface. Please enter the name of your primary network interface:

#

Device "#####" does not exist. 127.0.0.: error fetching interface information: Device not found

[ERROR] Fatal Error Occurred: Unable to determine IP address of selected interface. Cannot continue. Cannot continue installation. vm6-install.sh: 340: vm6-install.sh: cannot create /tmp/.virtualmin-21286/sources.list: Directory nonexistent mv: cannot stat '/tmp/.virtualmin-21286/sources.list': No such file or directory [ERROR] If you are unsure of what went wrong, you may wish to review the log [ERROR] in /root/virtualmin-install.log

[2018-09-18 00:04:39 EDT] [DEBUG] Cleaning up temporary files in /tmp/.virtualmin-21286. [2018-09-18 00:04:39 EDT] [WARNING] Unable to determine IP address of primary interface. [2018-09-18 00:09:08 EDT] [ERROR] Fatal Error Occurred: Unable to determine IP address of selected interface. Cannot continue. [2018-09-18 00:09:08 EDT] [ERROR] If you are unsure of what went wrong, you may wish to review the log [2018-09-18 00:09:08 EDT] [ERROR] in /root/virtualmin-install.log

I tried on two different VPS's on different providers, didn't work. I also tried using the IP, venet0, and venet0:0, and all have resulted in the same error.

Any help or suggestions will be helpful. I have root access to the VPS, and can do any suggested edit.

Thank You in advance.

M.S.

Tue, 11/20/2018 - 14:49
lexo_ch Pro Licensee

Same problem here. Tried to reinstall Debian 9.6 from scratch, rename interface to eth0, install vmware-tools right after fresh install (since it's running in a VM). We bought the professional full-version, reinstalled again, retried, changed eth hardware setup in ESXi -> Always same error:

[ERROR] Fatal Error Occurred: Unable to determine IP address of selected interface. Cannot continue

And then the script terminates and the network link is dead.

We suppose that the Debian netinst does not provide a tool/software that is required for the script to determine some network preferences. We installed the net tools (apt install net-tools) but this did not solve the issue either.

UPDATE - SOLVED

Reinstalled again. The problem was that the install.sh script has always disabled the network interface and then could not finish properly. This time we switched to second console (CTRL+ALT+F2), logged in again and simply issued

ifup eth0

The network interface was back online. Next we switched back to the primary console (CTRL+ALT+F1). The script was asking for the network interface name. We entered "eth0" and voilĂ , it worked out.

This is obviously a bug in the script and should be fixed.

Mon, 11/26/2018 - 23:53
Joe Pro Licensee
Joe's picture

Can you post the output of the following command:

# ip ro ls

That's how the script finds the primary interface...it lists the routes, grepping out the default route, and then doing some magic on that to get the name of the interface. Something must be weird with the interfaces for it to not be able to find it.

--

Check out the forum guidelines!

Tue, 11/27/2018 - 03:41
lexo_ch Pro Licensee
default via 185.104.85.129 dev eth0 onlink
185.104.85.128/27 dev eth0 proto kernel scope link src 185.104.85.140

Not sure if that helps though. I believe that one of the packages that is being installed by the script is causing the network link to go down.

Thu, 11/29/2018 - 08:24
MichaelS

Thank You to everyone that commented.

I will test what's been mentioned to also confirm the fix mentioned.

"The problem was that the install.sh script has always disabled the network interface"

Was this always the case, and if so, I wonder what changed since this was never an issue for about 6 or 7 years?

Thu, 11/29/2018 - 11:01
lexo_ch Pro Licensee

Well I completly re-installed the system from scratch 4 times in total: 2x 9.5 and 2x 9.6. For 9.6 I created a snapshot after second install and repeated the install.sh process twice. Second time I did what I wrote above (manually IFUP eth0 on second console) which solved the issue. Hope that helps!

Sun, 12/02/2018 - 01:22
palomnik Pro Licensee

I am also having this issue on a new Ramnode VPS (openvz). I have tried entering the ip, venet0 and venet0:0. This is on Ubuntu 16.04.

Installing as root using: bash install.sh --bundle LEMP --minimal

"ip ro ls" yields: default dev venet0 scope link

I get the following errors:

Device "link" does not exist. [WARNING] Unable to determine IP address of primary interface. Please enter the name of your primary network interface: venet0 Device "venet0" does not exist. rror fetching interface information: Device not found

[ERROR] Fatal Error Occurred: Unable to determine IP address of selected interface. Cannot continue. Cannot continue installation. install.sh: line 340: /tmp/.virtualmin-16357/sources.list: No such file or directory mv: cannot stat '/tmp/.virtualmin-16357/sources.list': No such file or directory [ERROR] If you are unsure of what went wrong, you may wish to review the log [ERROR] in /root/virtualmin-install.log

tailing the virtualimin-install.log gives me:

[2018/12/02 02:16:51] [INFO] - Configuring Virtualmin [2018/12/02 02:16:51] [INFO] - Succeeded [2018-12-02 02:16:51 EST] [DEBUG] Cleaning up temporary files in /tmp/.virtualmin-16357. [2018-12-02 02:16:51 EST] [WARNING] Unable to determine IP address of primary interface. [2018-12-02 02:17:04 EST] [ERROR] Fatal Error Occurred: Unable to determine IP address of selected interface. Cannot continue. [2018-12-02 02:17:04 EST] [ERROR] If you are unsure of what went wrong, you may wish to review the log [2018-12-02 02:17:04 EST] [ERROR] in /root/virtualmin-install.log

Last week I did the same thing on Linode with no problems, and in a virtualbox vm with ubuntu with no problems. What is special about this environment that makes it fail?