Centos 7: Error "Device eth1 does not seem to be present" with private network

[jong99]

Trying to use a private network with the new Centos 7 config doesn't work and results in the error below:

Steps to reproduce:

veewee vbox define 'centos70' CentOS-7.0-1406-x86_64-netinstall --force
veewee vbox build 'centos70' 
veewee vbox export 'centos70'
vagrant box add 'centos70' centos70.box
vagrant init 'centos70'

Now edit the Vagrantfile and uncomment the line:

  config.vm.network "private_network", ip: "192.168.33.10"

Then run:

vagrant up

This results in the following error:

==> default: Configuring and enabling network interfaces...
The following SSH command responded with a non-zero exit status.
Vagrant assumes that this means the command failed!

ARPCHECK=no /sbin/ifup eth1 2> /dev/null

Stdout from the command:

ERROR    : [/etc/sysconfig/network-scripts/ifup-eth] Device eth1 does not seem to be present, delaying initialization.


Stderr from the command:

[jong99]

Lots of discussion here on similar issues but nothing helps for Centos 7
hashicorp/vagrant#1777

[jong99]

Eventually managed to get it to work by changing the Vagrantfile line to

config.vm.network "private_network", ip: "192.168.33.10", auto_config: false

Would be nice if this could work straight out of the box though!

[mpapis]

should we update docs for this?

[hlangeveld]

See also mitchell/vagrant#4195

[mpapis]

so does updating vagrant to latest version help with this problem? - always update your software :)

[jgough]

Tried to reproduce with latest vagrant version but having issues with Guest Additions and can't get anything to work at the moment. Will let you know if I make progress.

[solomongifford]

I can confirm that upgrading to vagrant >=1.6.5 fixes the issue.

[mpapis]

thanks for confirming

[geerlingguy]

Another confirmation that simply upgrading to 1.6.5+ fixed the issue (I was running 1.6.3).

[ghost]

I am running 1.7.1 and I am getting this issue. Was this reintroduced with all the issues in 1.7.0?

[mpapis]

@monomidev please test with version lower then 1.7.0 to be sure - if it fails again please comment on hashicorp/vagrant#4195 or hashicorp/vagrant#1777 - whichever looks more similar

[ghost]

Sorry for the confusion but it looks like this is not caused by Vagrant. I was running into a couple of issues. When I first encountered this, I was getting an error with device eth1. I noticed there was a ifcfg-eth1 file but eth2 was the device vagrant was adding. It would then error on eth1 with the error mentioned in this thread. I found the following article on a similar issue with ubuntu and vagrant:

http://able.cd/b/2012/04/09/vagrant-broken-networking-when-packaging-ubuntu-boxes/

Quite old but was the case me. When I remove the rules file, vagrant was having the same error for eth2 and then eth3 was the actual device. I noticed the ifcfg-eth1 file was still present. The fix I did:

1. In virtualbox, I removed all but one adapter.
2. I removed the ifcfg-eth1 file.
3. I removed the rules file "/etc/udev/rules.d/70-persistent-net.rules"

Now vagrant attempts to create device eth1 and writes a config file for eth1 when setting a private_network.

[mpapis]

are you saying that automated removal of /etc/udev/rules.d/70-persistent-net.rules during one of the steps would solve the problem? can you open pull request that automates it?

[klub]

Removing /etc/udev/rules.d/70-persistent-net.rules
seems to have worked for me too.

[ghost]

@mpapis sorry for the delay, I think it is a combination of bad base box configuration (causing the extra ifcfg file) and the rules file. I will see what I can put together.

[dankeder]

I have vagrant 1.7.2 and I am encoutering this issue too.

As far as I can tell, the problem is that Vagrant expects the second network interface to be called eth1, and it wants to create configuration for it in /etc/sysconfig/network-scripts/ifcfg-eth1. But network interfaces have a different naming convention in Centos 7 (and Fedora as well) - the network interface in question is actually called like enp0s8 (or something similar - you can get the real interface name by running ip link in the VM shell).

The solution would be to explicitly specify the interface name in Vagrantfile - which I don't know how to do, but maybe somebody else does.

A possible workaround which I used is to disable interface renaming by adding net.ifnames=0 to GRUB_CMDLINE_LINUX=... line in /etc/sysconfig/grub in the base VM image, and running grub2-mkconfig > /etc/grub2.cfg.

[mpapis]

I think net.ifnames=0 is the best option, we actually want the old behavior for virtual machines, the new names are only trouble in our use case - btw. I think I used the same thing for openSUSE

[Stono]

So I have this issue (I think) with RHEL7.
I have created an image using net.ifnames=0 etc so my interfaces are named correctly, however when I add a private network using Vagrant I get:

The following SSH command responded with a non-zero exit status.
Vagrant assumes that this means the command failed!

/sbin/ifdown 

Stdout from the command:



usage: ifdown <device name>

When i log onto a box there is a random file....

/etc/sysconfig/network-scripts/ifcfg-

Karl

[mpapis]

You will have to debug this process yourself, I'm stepping down to only advise on fixing the issues (see: http://niczsoft.com/2015/02/changes-in-my-open-source-life/)

[pebreo]

I was able to fix this issue by upgrading to Vagrant v1.7.2.

[pgporada]

This still happens to me on Vagrant v1.7.2

[anhcuong]

the problem happens on Vagrant v1.7.2 !!!

[DanielAron]

Same problem here following this tutorial:
https://github.com/openMF/mifosx/wiki/MifosX-Vagrant-Box-Setup
Vagrant v1.7.2