1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306
|
/*
Copyright The containerd Authors.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
*/
package main
import (
"context"
"flag"
"fmt"
"net"
"os"
"runtime"
"strings"
"github.com/containernetworking/plugins/pkg/ns"
"github.com/sirupsen/logrus"
"github.com/vishvananda/netlink"
"sigs.k8s.io/yaml"
"github.com/containerd/nri/pkg/api"
"github.com/containerd/nri/pkg/stub"
)
const (
// Prefix of the key used for network device annotations.
netdeviceKey = "netdevices.nri.containerd.io"
)
var (
log *logrus.Logger
verbose bool
)
// an annotated netdevice
// https://man7.org/linux/man-pages/man7/netdevice.7.html
type netdevice struct {
Name string `json:"name"` // name in the runtime namespace
NewName string `json:"new_name"` // name inside the pod namespace
Address string `json:"address"`
Prefix int `json:"prefix"`
MTU int `json:"mtu"`
}
func (n *netdevice) inject(nsPath string) error {
// Lock the OS Thread so we don't accidentally switch namespaces
runtime.LockOSThread()
defer runtime.UnlockOSThread()
containerNs, err := ns.GetNS(nsPath)
if err != nil {
return err
}
defer containerNs.Close()
hostDev, err := netlink.LinkByName(n.Name)
if err != nil {
return err
}
_, err = moveLinkIn(hostDev, containerNs, n.NewName)
if err != nil {
return fmt.Errorf("failed to move link %v", err)
}
return nil
}
// remove the network device from the Pod namespace and recover its name
// Leaves the interface in down state to avoid issues with the root network.
func (n *netdevice) release(nsPath string) error {
// Lock the OS Thread so we don't accidentally switch namespaces
runtime.LockOSThread()
defer runtime.UnlockOSThread()
containerNs, err := ns.GetNS(nsPath)
if err != nil {
return err
}
defer containerNs.Close()
err = moveLinkOut(containerNs, n.NewName)
if err != nil {
return err
}
return nil
}
// our injector plugin
type plugin struct {
stub stub.Stub
}
func (p *plugin) RunPodSandbox(_ context.Context, pod *api.PodSandbox) error {
log.WithField("namespace", pod.GetNamespace()).WithField("name", pod.GetName).Debug("Started pod...")
if verbose {
dump("RunPodSandbox", "pod", pod)
}
// inject associated netdevices (based on received pod annotations) into the pod
// network namespace that will be attached to the pod's containers
netdevices, err := parseNetdevices(pod.Annotations)
if err != nil {
return err
}
if len(netdevices) == 0 {
return nil
}
// get the pod network namespace
var ns string
for _, namespace := range pod.Linux.GetNamespaces() {
if namespace.Type == "network" {
ns = namespace.Path
break
}
}
// Pods running on the host network namespace has this value empty
if ns == "" {
log.WithField("namespace", pod.GetNamespace()).WithField("name", pod.GetName).Info("Pod using host namespace, skipping ...")
return fmt.Errorf("trying to inject network device on host network Pod")
}
// attach the network devices to the pod namespace
for _, n := range netdevices {
err = n.inject(ns)
if err != nil {
return nil
}
}
return nil
}
func (p *plugin) StopPodSandbox(_ context.Context, pod *api.PodSandbox) error {
log.WithField("namespace", pod.GetNamespace()).WithField("name", pod.GetName).Debug("Stopped pod...")
if verbose {
dump("StopPodSandbox", "pod", pod)
}
// release associated devices of the netdevice to the Pod
netdevices, err := parseNetdevices(pod.Annotations)
if err != nil {
return err
}
if len(netdevices) == 0 {
return nil
}
// get the pod network namespace
var ns string
for _, namespace := range pod.Linux.GetNamespaces() {
if namespace.Type == "network" {
ns = namespace.Path
break
}
}
// TODO check host network namespace
if ns == "" {
return nil
}
// release the network devices from the pod namespace
for _, n := range netdevices {
err = n.release(ns)
if err != nil {
return nil
}
}
return nil
}
func parseNetdevices(annotations map[string]string) ([]netdevice, error) {
var (
key string
annotation []byte
netdevices []netdevice
)
// look up effective device annotation and unmarshal devices
for _, key = range []string{
netdeviceKey + "/pod",
netdeviceKey,
} {
if value, ok := annotations[key]; ok {
annotation = []byte(value)
break
}
}
if annotation == nil {
return nil, nil
}
if err := yaml.Unmarshal(annotation, &netdevices); err != nil {
return nil, fmt.Errorf("invalid device annotation %q: %w", key, err)
}
// validate and default
for _, n := range netdevices {
if n.NewName == "" {
n.NewName = n.Name
}
if n.Address != "" {
ip := net.ParseIP(n.Address)
if ip == nil {
return nil, fmt.Errorf("error parsing address %s", n.Address)
}
if n.Prefix == 0 {
if ip.To4() == nil {
n.Prefix = 128
} else {
n.Prefix = 32
}
}
}
}
return netdevices, nil
}
// Dump one or more objects, with an optional global prefix and per-object tags.
func dump(args ...interface{}) {
var (
prefix string
idx int
)
if len(args) == 1 {
prefix = args[0].(string)
idx++
}
for ; idx < len(args)-1; idx += 2 {
tag, obj := args[idx], args[idx+1]
msg, err := yaml.Marshal(obj)
if err != nil {
log.Infof("%s: %s: failed to dump object: %v", prefix, tag, err)
continue
}
if prefix != "" {
log.Infof("%s: %s:", prefix, tag)
for _, line := range strings.Split(strings.TrimSpace(string(msg)), "\n") {
log.Infof("%s: %s", prefix, line)
}
} else {
log.Infof("%s:", tag)
for _, line := range strings.Split(strings.TrimSpace(string(msg)), "\n") {
log.Infof(" %s", line)
}
}
}
}
func main() {
var (
pluginName string
pluginIdx string
opts []stub.Option
err error
)
log = logrus.StandardLogger()
log.SetFormatter(&logrus.TextFormatter{
PadLevelText: true,
})
flag.StringVar(&pluginName, "name", "", "plugin name to register to NRI")
flag.StringVar(&pluginIdx, "idx", "", "plugin index to register to NRI")
flag.BoolVar(&verbose, "verbose", false, "enable (more) verbose logging")
flag.Parse()
if pluginName != "" {
opts = append(opts, stub.WithPluginName(pluginName))
}
if pluginIdx != "" {
opts = append(opts, stub.WithPluginIdx(pluginIdx))
}
p := &plugin{}
if p.stub, err = stub.New(p, opts...); err != nil {
log.Fatalf("failed to create plugin stub: %v", err)
}
err = p.stub.Run(context.Background())
if err != nil {
log.Errorf("plugin exited with error %v", err)
os.Exit(1)
}
}
|