networking - 在带有 Terraform/KVM 的 CentOS 7 上使用 Cloud-Init 设置静态 IP

Question

我使用 Terraform 和 KVM 创建了一个 CentOS 7 VM。作为测试服务器，这个虚拟机应该有一个静态 IP，应该可以使用 Cloud-Init :

version: 2
ethernets:
  eth0:
    # match:
    #   name: eth0
    dhcp4: false
    dhcp6: false
    addresses: 
      - 10.18.3.2/24
    gateway4: 10.18.3.1

现在我们重新创建虚拟机并打开它的控制台:

terraform destroy -auto-approve
terraform apply -auto-approve
terraform console centos

这会从 Cloud-Init 生成以下输出

[   10.073544] cloud-init[626]: Cloud-init v. 18.5 running 'init-local' at Wed, 15 Jan 2020 17:34:43 +0000. Up 10.04 seconds.
[  OK  ] Started Initial cloud-init job (pre-networking).
[  OK  ] Reached target Network (Pre).
         Starting LSB: Bring up/down networking...
[FAILED] Failed to start LSB: Bring up/down networking.
See 'systemctl status network.service' for details.
         Starting Initial cloud-init job (metadata service crawler)...
[  OK  ] Reached target Network.
         Starting Postfix Mail Transport Agent...
         Starting Dynamic System Tuning Daemon...
[  OK  ] Started Dynamic System Tuning Daemon.
[  310.701064] cloud-init[820]: Cloud-init v. 18.5 running 'init' at Wed, 15 Jan 2020 17:39:44 +0000. Up 310.67 seconds.
[  310.724568] cloud-init[820]: ci-info: +++++++++++++++++++++++++++++++++++Net device info+++++++++++++++++++++++++++++++++++
[  310.727151] cloud-init[820]: ci-info: +--------+------+---------------------------+-----------+-------+-------------------+
[  310.729810] cloud-init[820]: ci-info: | Device |  Up  |          Address          |    Mask   | Scope |     Hw-Address    |
[  310.732602] cloud-init[820]: ci-info: +--------+------+---------------------------+-----------+-------+-------------------+
[  310.735405] cloud-init[820]: ci-info: |  eth0  | True | fe80::5054:ff:fe6a:ca6/64 |     .     |  link | 52:54:00:6a:0c:a6 |
[  310.738253] cloud-init[820]: ci-info: |   lo   | True |         127.0.0.1         | 255.0.0.0 |  host |         .         |
[  310.740787] cloud-init[820]: ci-info: |   lo   | True |          ::1/128          |     .     |  host |         .         |
[  310.743060] cloud-init[820]: ci-info: +--------+------+---------------------------+-----------+-------+-------------------+

在 [FAILED] 线路上花了大约 5 分钟无法启动 LSB:启动/关闭网络失败。正如我们所见，机器没有按照 Cloud-Init 的要求分配 ipc4 地址。

为什么我的静态 IP 地址没有分配？我还尝试将 ens3 作为接口(interface)名称，但没有成功。

Terraforms test.tf 文件

provider "libvirt" {
  uri = "qemu:///system"
}
resource "libvirt_pool" "test_pool" {
  name = "test_pool"
  type = "dir"
  path = "/tmp/kvm_test"
}
resource "libvirt_volume" "centos7-img" {
  name    = "centos7.qcow2"
  pool    = libvirt_pool.test_pool.name
  source =  "/var/lib/libvirt/images/CentOS-7-x86_64-GenericCloud.qcow2"
  format =  "qcow2"
}
data "template_file" "cloudinit_data" {
  template = file("cloudinit.cfg")
}
data "template_file" "cloudinit_network" {
  template = file("network.cfg")
}
resource "libvirt_cloudinit_disk" "cloudinit" {
  name           = "cloudinit.iso"
  # https://github.com/hashicorp/terraform/issues/7919#issuecomment-320816276
  user_data      = data.template_file.cloudinit_data.rendered
  network_config = data.template_file.cloudinit_network.rendered
  pool           = libvirt_pool.test_pool.name
}

resource "libvirt_network" "test_network" {
   name = "test_network"
   addresses = ["10.18.3.0/24"]
   dhcp {
      enabled = false
   }
}
resource "libvirt_domain" "centos" {
  name   = "centos"
  memory = "1024"
  vcpu   = 4
  cloudinit = libvirt_cloudinit_disk.cloudinit.id

  network_interface {
    #network_name = "default"
    network_id = libvirt_network.test_network.id
  }

  disk {
    volume_id = libvirt_volume.centos7-img.id
  }

  console {
    type = "pty"
    target_type = "serial"
    target_port = "0"
  }
  console {
    type        = "pty"
    target_type = "virtio"
    target_port = "1"
  }

  graphics {
    type = "spice"
    listen_type = "address"
    autoport = true
  }
}

cloudinit.cfg

#cloud-config
# https://cloudinit.readthedocs.io/en/latest/topics/modules.html
timezone: Europe/Berlin

fqdn: myhost.internal
manage_etc_hosts: true
resize_rootfs: true

ssh_authorized_keys:
  - ssh-rsa mykey

ssh_pwauth: true
password: password
chpasswd:
  list: |
    root:password
    centos:password
  expire: false

# Here I set the yum proxy and update all packages
# runcmd:
# ...

网络服务日志

# systemctl status network.service -l
● network.service - LSB: Bring up/down networking
   Loaded: loaded (/etc/rc.d/init.d/network; bad; vendor preset: disabled)
   Active: failed (Result: timeout) since Wed 2020-01-15 18:39:44 CET; 3min 11s ago
     Docs: man:systemd-sysv-generator(8)
  Process: 654 ExecStart=/etc/rc.d/init.d/network start (code=killed, signal=TERM)
   CGroup: /system.slice/network.service
           └─1000 /sbin/dhclient -q -lf /var/lib/dhclient/dhclient--eth0.lease -pf /var/run/dhclient-eth0.pid -H myhost eth0

Jan 15 18:40:50 myhost dhclient[1000]: DHCPDISCOVER on eth0 to 255.255.255.255 port 67 interval 6 (xid=0x60f0f449)
Jan 15 18:40:56 myhost dhclient[1000]: DHCPDISCOVER on eth0 to 255.255.255.255 port 67 interval 6 (xid=0x60f0f449)
Jan 15 18:41:02 myhost dhclient[1000]: DHCPDISCOVER on eth0 to 255.255.255.255 port 67 interval 11 (xid=0x60f0f449)
Jan 15 18:41:13 myhost dhclient[1000]: DHCPDISCOVER on eth0 to 255.255.255.255 port 67 interval 20 (xid=0x60f0f449)
Jan 15 18:41:33 myhost dhclient[1000]: DHCPDISCOVER on eth0 to 255.255.255.255 port 67 interval 20 (xid=0x60f0f449)
Jan 15 18:41:53 myhost dhclient[1000]: DHCPDISCOVER on eth0 to 255.255.255.255 port 67 interval 17 (xid=0x60f0f449)
Jan 15 18:42:10 myhost dhclient[1000]: DHCPDISCOVER on eth0 to 255.255.255.255 port 67 interval 17 (xid=0x60f0f449)
Jan 15 18:42:27 myhost dhclient[1000]: DHCPDISCOVER on eth0 to 255.255.255.255 port 67 interval 13 (xid=0x60f0f449)
Jan 15 18:42:40 myhost dhclient[1000]: DHCPDISCOVER on eth0 to 255.255.255.255 port 67 interval 9 (xid=0x60f0f449)
Jan 15 18:42:49 myhost dhclient[1000]: DHCPDISCOVER on eth0 to 255.255.255.255 port 67 interval 17 (xid=0x60f0f449)

解决方法:不要使用 Cloud-Init 进行网络配置

目前我发现的唯一解决方法是禁用 Cloud-Init 的网络部分，这意味着不渲染 network.cfg 文件。相反，在 network_interface 部分设置静态 IP 有效:

resource "libvirt_domain" "centos" {
  name   = "centos"
  memory = "1024"
  vcpu   = 4
  cloudinit = libvirt_cloudinit_disk.cloudinit.id

  network_interface {
    network_id = libvirt_network.test_network.id

    hostname  = "centos"
    addresses = ["10.18.3.2"]
  }
  # ...
}

Answer 1

我遇到了同样的场景，简单地在配置文件中指定 dhcp4 或 dhcp6 会导致虚拟机尝试通过 dhcp 提取地址(即使值设置为 false。)

对我来说，解决方法是将 dhcp4 和 dhcp6 排除在配置文件之外，并且在快速启动和静态网络配置到位的情况下，一切都按预期工作。