CentOS - May 2017 - rpcbind fails to start after creating virbr0 bridge

Hi,

In my office I have an HP Proliant server running CentOS 7. It's a
minimal install without GUI. The machine has two network interface
cards, and for now it's acting as gateway/firewall, and it's running
ntpd, Dnsmasq and Rsnapshot.

I'd like this machine to also be a KVM virtualization host, so I
installed qemu-kvm and libvirt. The interface facing the LAN is enp3s0,
so I created a virbr0 bridge like this.

  # /etc/sysconfig/network-scripts/ifcfg-enp3s0
  DEVICE=enp3s0
  TYPE=Ethernet
  ONBOOT=yes
  BRIDGE=virbr0

And then:

  # /etc/sysconfig/network-scripts/ifcfg-virbr0
  DEVICE=virbr0
  TYPE=Bridge
  ONBOOT=yes
  BOOTPROTO=static
  IPADDR=192.168.2.1
  NETMASK=255.255.255.0

I replaced enp3s0 in my firewall script and in dnsmasq.conf and
rebooted, and everything seems OK now... or almost. Here's an unexpected
error I get.

  # systemctl --failed
    UNIT           LOAD   ACTIVE SUB    DESCRIPTION
  ? rpcbind.socket loaded failed failed RPCbind Server Activation Socket

This error appeared since I created my virbr0 interface, so I presume it
has something to do with it. Unfortunately, I'm clueless.

Any suggestions?

Niki

-- 
Microlinux - Solutions informatiques durables
7, place de l'?glise - 30730 Montpezat
Web  : http://www.microlinux.fr
Mail : info at microlinux.fr
T?l. : 04 66 63 10 32

On 05/09/2017 04:08 AM, Nicolas Kovacs wrote:
> I'd like this machine to also be a KVM virtualization host, so I
> installed qemu-kvm and libvirt. The interface facing the LAN is enp3s0,
> so I created a virbr0 bridge like this.

virbr0 is a device managed by libvirt, and it'll come up later than the 
devices managed by the network or NetworkManager services.  Don't 
enslave your ethernet devices to it.

Instead, create a new bridge.  Revert your changes and run:

virsh iface-bridge enp3s0 br0 --no-stp