Libvirt Resume Guest Startup Issues CentOS 7

Home » CentOS » Libvirt Resume Guest Startup Issues CentOS 7
CentOS 6 Comments

Here is the relevant log snippet:

journal: libvirt version: 1.2.8, package: 16.el7_1.3 (CentOS BuildSystem
, 2015-05-12-20:12:58, worker1.bsys.CentOS.org)

journal: failed to connect to monitor socket: No such process

journal: internal error: process exited while connecting to monitor:
((null):1937): Spice-Warning **: reds.c:3036:reds_init_socket:
getaddrinfo(127.0.0.1,5900): Address family for hostname not supported

qemukvm: failed to initialize spice server

If I remove graphics devices from the VM it saves/resumes properly on host restart.

https://bugzilla.redhat.com/show_bug.cgi?idr1350

I feel like this is an issue of libvirt starting before networking, but I am a fish out of water with systemd.

Any thoughts?

-Jason

6 thoughts on - Libvirt Resume Guest Startup Issues CentOS 7

  • Are you using the “network” or “NetworkManager” service to configure your network devices?

  • Likely, if that is how they are out of the box.

    # for i in /etc/sysconfig/network-scripts/ifcfg-*; do echo -e ‘\n#’ $i; cat
    $i; done

    # /etc/sysconfig/network-scripts/ifcfg-br0
    DEVICE=br0
    TYPE=Bridge BOOTPROTO=dhcp DEFROUTE=yes PEERDNS=yes PEERROUTES=yes IPV4_FAILURE_FATAL=no IPV6INIT=no IPV6_AUTOCONF=no IPV6_DEFROUTE=no IPV6_PEERDNS=no IPV6_PEERROUTES=no IPV6_FAILURE_FATAL=no ONBOOT=yes

    # /etc/sysconfig/network-scripts/ifcfg-eno1
    #TYPE=Ethernet NAME=eno1
    UUID

  • As a desperate attempt, I disabled IPv6
    [https://wiki.CentOS.org/FAQ/CentOS7#head-8984faf811faccca74c7bcdd74de7467f2
    fcd8ee]. The VMs start properly, but now I am with out IPv6. So this is a workaround for the interim.

    I should have been more clear, I made NO customizations except as pasted previously. What ever the minimal install uses is what I have.

    I really do not know enough about systemd, but I think this might answer your question.

    # systemctl | grep -i net sys-devices-pci0000:00-0000:00:09.0-0000:01:00.0-0000:02:01.0-0000:03:00.0-0
    000:04:00.0-net-eno1.device loaded active plugged PowerEdge 2970 Broadcom NetXtreme II BCM5708
    sys-devices-pci0000:00-0000:00:09.0-0000:01:00.0-0000:02:02.0-0000:05:00.0-0
    000:06:00.0-net-eno2.device loaded active plugged PowerEdge 2970 Broadcom NetXtreme II BCM5708
    sys-devices-virtual-net-br0.device loaded active plugged /sys/devices/virtual/net/br0
    sys-devices-virtual-net-vnet0.device loaded active plugged /sys/devices/virtual/net/vnet0
    sys-subsystem-net-devices-br0.device loaded active plugged /sys/subsystem/net/devices/br0
    sys-subsystem-net-devices-eno1.device loaded active plugged PowerEdge 2970 Broadcom NetXtreme II BCM5708
    sys-subsystem-net-devices-eno2.device loaded active plugged PowerEdge 2970 Broadcom NetXtreme II BCM5708
    sys-subsystem-net-devices-vnet0.device loaded active plugged /sys/subsystem/net/devices/vnet0
    netcf-transaction.service loaded active exited Rollback uncommitted netcf network config change transactions network.service loaded active exited LSB: Bring up/down networking NetworkManager-wait-online.service loaded failed failed Network Manager Wait Online NetworkManager.service loaded active running Network Manager rhel-import-state.service loaded active exited Import network configuration from initramfs network-online.target loaded active active Network is Online network.target loaded active active Network

    -Jason

  • I’ve looked at this a bit… But, bear in mind that I don’t use Spice, so everything I say may be wrong.

    I don’t think libvirtd is starting before networking, regardless of the service. However, if you’d like to test that theory, you can copy
    /usr/lib/systemd/system/libvirtd.service to /etc/systemd/system and edit it. Replace the line “After=network.target” with
    “After=network-online.target”, then reboot. See if that affects the problem in any way.

    The problem looks vaguely like the server is trying to use an IPv4
    address (127.0.0.1) and an IPv6 address family. I’m not sure why it would do that. Were there any specific directions you followed to set up Spice?

    I don’t think this bug is relevant, as it results in a return value of EAI_NONAME. In your case, getaddrinfo returns EAI_FAMILY.

  • Same *exact* error with VNC.

    Will try, but not until Wednesday / Thursday PM.

    Dell PowerEdge 2970
    ====================Install CentOS 7 x64 minimum.
    [left as default]
    vi /etc/sysconfig/network-scripts/ifcfg-eno1
    [change onboot=yes]
    service network restart yum update -y && yum install nano logwatch ntpd ntpdate && reboot nano /etc/hostname yum group install ‘Virtualization Host’
    chkconfig libvirt-guests on yum install virt-manager xauth nano /etc/sysconfig/network-scripts/ifcfg-br0
    [delete auto created virtbr0]
    Reboot
    [create vm with defaults – CentOS 7 x64 minimal]
    virsh autostart CentOS7.0
    [reboot]
    [debug, pull out hair, email CentOS list, remove ipv6, remain puzzled]

    This and its linked tickets is where I figured to disable IPv6.