【katacontainer基础知识】katacontainer介绍与原理

WHY ?

       弥补了传统容器技术安全性的缺点，Kata Containers通过使用硬件虚拟化来达到容器隔离的目的。每一个container/pod 都是基于一个独立的kernel实例来作为一个轻量级的虚拟机。自从每一个container/pod运行与独立的虚拟机上，他们不再从宿主机内核上获取相应所有的权限。

WHAT ?

      kata containers是由OpenStack基金会管理的容器项目。kata containers整合了Intel的 Clear Containers 和 Hyper.sh 的 runV，能够支持不同平台的硬件，并符合Open Container Initiative规范，同时还可以兼容k8s的 CRI（Container Runtime Interface）接口规范。项目包含几个配套组件，即Runtime，Agent， Proxy，Shim等。项目已于6月份release了1.0版本。
      从docker架构上看，kata-container和原来的runc是平级的。大家知道docker只是管理容器生命周期的框架，真正启动容器最早用的是LXC，然后是runc，现在也可以换成kata了。所以说kata-container可以当做docker的一个插件，启动kata-container可以通过docker命令。

kata container组件

   1. Agent

       Kata-agent运行在guest负责管理容器。Kata-agent的执行单元是定义了一系列命名空间的沙盒。每个VM可以运行多个容器，支持k8s一个pod运行多个容器的需求。不过目前docker中，kata-runtime只能一个pod一个容器。Kata-agent通过gRPC和其他kata组件通信。Kata-agent使用libcontainer管理容器的生命周期

  2. Runtime

     kata-runtime是一个OCI兼容的容器运行时，负责处理OCI运行时规范指定的所有命令并启动kata-shim实例

    配置文件是/usr/share/defaults/kata-containers/configuration.toml

# XXX: WARNING: this file is auto-generated.
# XXX:
# XXX: Source file: "cli/config/configuration.toml.in"
# XXX: Project:
# XXX: Name: Kata Containers
# XXX: Type: kata[hypervisor.qemu]
path = "/usr/bin/qemu-lite-system-x86_64"
kernel = "/usr/share/kata-containers/vmlinuz.container"
image = "/usr/share/kata-containers/kata-containers.img"
machine_type = "pc"# Optional space-separated list of options to pass to the guest kernel.
# For example, use `kernel_params = "vsyscall=emulate"` if you are having
# trouble running pre-2.15 glibc.
#
# WARNING: - any parameter specified here will take priority over the default
# parameter value of the same name used to start the virtual machine.
# Do not set values here unless you understand the impact of doing so as you
# may stop the virtual machine from booting.
# To see the list of default parameters, enable hypervisor debug, create a
# container and look for &＃39;default-kernel-parameters&＃39; log entries.
kernel_params = ""# Path to the firmware.
# If you want that qemu uses the default firmware leave this option empty
firmware = ""

  3. Proxy

        默认使用virtio-serial和VM通信。VM可以运行多个容器进程。在使用virtio-serial的情况下，与每个进程相关联的I/O流需要在主机上多路复用和解复用。
        Kata-proxy给多个kata-shim和kata-runtime客户端提供对kata-agent提供访问，它的主要作用是在每个kata-shim和kata-agent之间路由I/O流和信号。Kata-proxy连接到kata-agent的unix域套接字上，这个套接字是kata-proxy启动时kata-runtime提供的

  4. Shim

        runtime运行在宿主机上，不能直接监控运行在虚拟机里的进程，最多只能看到QEMU进程。kata-shim监控容器进程，处理容器的所有I/O流，包括stdout、stdin和stderr，以及转发所有的要发送出去的信号。
        Kata-shim还有其他功能：

• 通过一个UNIX域套接字连接到kata-proxy。这个套接字在kata-runtime启动kata-shim的时候，由kata-runtime传给kata-shim，同时带上了containerID和execID，两个ID用来识别shim管理的是哪个容器。
• 读取VM内部容器进程的输出流和错误流
• 使用SignalProcessRequest API转发从reaper到kata-proxy的信号
• 监控终端修改，并使用grpc TtyWinResize API转发到kata-proxy

kata-runtime kata-env

[Meta]
  Version = "1.0.18"

[Runtime]
  Debug = false
  Path = "/usr/bin/kata-runtime"
  [Runtime.Version]
    Semver = "1.3.1"
    Commit = "258eae0"
    OCI = "1.0.1"
  [Runtime.Config]
    Path = "/etc/kata-containers/configuration.toml"

[Hypervisor]
  MachineType = "pc"
  Version = "QEMU emulator version 2.11.0\nCopyright (c) 2003-2017 Fabrice Bellard and the QEMU Project developers"
  Path = "/usr/bin/qemu-lite-system-x86_64"
  BlockDeviceDriver = "virtio-scsi"
  EntropySource = "/dev/urandom"
  Msize9p = 8192
  MemorySlots = 10
  Debug = false
  UseVSock = false

[Image]
  Path = "/usr/share/kata-containers/kata-containers-image_clearlinux_1.3.1_agent_c7fdd324cda.img"

[Kernel]
  Path = "/usr/share/kata-containers/vmlinuz-4.14.67.16-4.4.container"
  Parameters = ""

[Initrd]
  Path = ""

[Proxy]
  Type = "kataProxy"
  Version = "kata-proxy version 1.3.1-d364b2e"
  Path = "/usr/libexec/kata-containers/kata-proxy"
  Debug = false

[Shim]
  Type = "kataShim"
  Version = "kata-shim version 1.3.1-58f757d"
  Path = "/usr/libexec/kata-containers/kata-shim"
  Debug = false

[Agent]
  Type = "kata"

[Host]
  Kernel = "3.10.0-957.el7.x86_64"
  Architecture = "amd64"
  VMCOntainerCapable= true
  SupportVSocks = false
  [Host.Distro]
    Name = "CentOS Linux"
    Version = "7"
  [Host.CPU]
    Vendor = "GenuineIntel"
    Model = "Intel(R) Core(TM) i7-6500U CPU @ 2.50GHz"

[Netmon]
  Version = "kata-netmon version 1.3.1"
  Path = "/usr/libexec/kata-containers/kata-netmon"
  Debug = false
  Enable = false

HOW ?

  1. centos 7 安装docker

yum -y install yum-utils
yum-config-manager --add-repo http://mirrors.aliyun.com/docker-ce/linux/centos/docker-ce.repo
yum -y install docker-ce

  2. centos 7 安装kata-containers

source /etc/os-release
echo $VERSION_ID
VERSION_ID=$VERSION_ID yum-config-manager --add-repo \
"http://download.opensuse.org/repositories/home:/katacontainers:/release/CentOS_${VERSION_ID}/home:katacontainers:release.repo"
yum -y install kata-runtime kata-proxy kata-shim

  3. 配置docker启动runtime参数

     --add-runtime kata-runtime=/usr/bin/kata-runtime --default-runtime=kata-runtime

[Unit]
Description=Docker Application Container Engine
Documentation=http://docs.docker.com
After=network.target docker-storage-setup.service
Wants=docker-storage-setup.service[Service]
Type=notify
EnvirOnment=GOTRACEBACK=crash
ExecReload=/bin/kill -s HUP $MAINPID
Delegate=yes
KillMode=process
ExecStart=/usr/bin/dockerd -s overlay2 --storage-opt overlay2.override_kernel_check=1 \--add-runtime kata-runtime=/usr/bin/kata-runtime --default-runtime=kata-runtime \$DOCKER_OPTS \$DOCKER_STORAGE_OPTIONS \$DOCKER_NETWORK_OPTIONS \$DOCKER_DNS_OPTIONS \$INSECURE_REGISTRY
LimitNOFILE=1048576
LimitNPROC=1048576
LimitCORE=infinity
TimeoutStartSec=1min
Restart=on-abnormal[Install]
WantedBy=multi-user.target

  4. 重启docker进程

     systemctl daemon-reload && systemctl restart docker

    启动容器，查看容器状态，docker stats，同时发现启动qemu进程

katacontainer kubernetes 集成安装

a. 使用 rpm 安装

     http://download.opensuse.org/repositories/home:/katacontainers:/releases:/x86_64:/master/CentOS_7/x86_64/
     http://download.opensuse.org/repositories/home:/katacontainers:/releases:/x86_64:/stable-1.3/CentOS_7/x86_64/4 安装

b. runc

   不用 docker 跟 k8s 通信而是用 cri-o，同时支持普通容器和 kata 容器，要单独安装 runc
   https://github.com/opencontainers/runc/releases/download/v1.0.0-rc8/runc.amd64

   chmod +x runc.amd64
   mv runc.amd64 /usr/bin/runc

c. 安装 crictl CLI

   https://github.com/kubernetes-sigs/cri-tools/releases/download/v1.13.0/crictl-v1.13.0-linux-amd64.tar.gz

d. crio

   https://kubernetes.io/docs/setup/production-environment/container-runtimes/#cri-o

   Prerequisites
modprobe overlay
modprobe br_netfilter

# Setup required sysctl params, these persist across reboots.
cat > /etc/sysctl.d/99-kubernetes-cri.conf < net.bridge.bridge-nf-call-iptables  = 1
net.ipv4.ip_forward                 = 1
net.bridge.bridge-nf-call-ip6tables = 1
EOF

sysctl --system

# Install prerequisites
yum-config-manager --add-repo=https://cbs.centos.org/repos/paas7-crio-311-candidate/x86_64/os/

# Install CRI-O
yum install --nogpgcheck cri-o

/etc/crio/crio.conf 可以使用命令 "crio config --default >/etc/crio/crio.conf" 创建，然后修改 registries 即可：

e. 配置 crio

修改 kata-runtime

# runtime_untrusted_workload is the OCI compatible runtime used for untrusted
# container workloads. This is an optional setting, except if
# default_container_trust is set to "untrusted".
runtime_untrusted_workload = "/usr/bin/kata-runtime"

修改镜像仓库

registries = [
    "docker.io"
]

修改 cgroupfs 等 

# cgroup_manager is the cgroup management implementation to be used
# for the runtime.
cgroup_manager = "cgroupfs"

修改 pause 地址

# pause_image is the image which we use to instantiate infra containers.
pause_image = "zhangzhonglin/pause-amd64:3.0"

修改 plugin_dir 路径

# plugin_dir is is where CNI plugin binaries are stored.
plugin_dir = "/opt/k8s/bin"
 

Start CRI-O  systemctl start crio

f. kubelet 配置

   添加参数 --container-runtime=remote --runtime-request-timeout=15m --container-runtime-endpoint=unix:///var/run/crio/crio.sock

测试启动一个 deployment 以注解 io.kubernetes.cri-o.TrustedSandbox: "false"：

spec:
  replicas: 1
  selector:
    matchLabels:
      app: wordpress-hostpath
      tier: mysql-hostpath
  strategy:
    type: Recreate
  template:
    metadata:
      annotations:
        io.kubernetes.cri-o.TrustedSandbox: "false"

crictl image

# crictl images
IMAGE                                 TAG                 IMAGE ID            SIZE
docker.io/calico/cni                  v3.7.4              203668d151b2e       139MB
docker.io/calico/kube-controllers     v3.7.4              e67ede28cc7e4       46.8MB
docker.io/calico/node                 v3.7.4              84b65b552a8f0       159MB
docker.io/coredns/coredns             1.5.0               7987f0908caf3       42.7MB
docker.io/zhangzhonglin/pause-amd64   3.0                 99e59f495ffaa       753kB

发现启动 crio common 进程

启动 qemu 进程

总结：

   这个镜像管理比较搓，下载到本地还是不可以，必须是 crio.cfg 配置中的仓库必须能通才可以，containerd 最起码可以管理镜像可以 load

  安装 kata-runtime，先验证环境是否支持虚拟化环境，否则就白玩了

   kata-runtime kata-env，中 VMCOntainerCapable= true 这个是支持虚拟化环境。

2. 使用 containerd 作为 CRI

下载地址：https://github.com/containerd/containerd/releases
解压至目录 /usr/bin 下sudo systemctl start containerdsudo systemctl status containerdcontainerd config default 生成默认配置，位置/etc/containerd/config.toml

     2.1 crictl配置

# cat /etc/crictl.yaml 
runtime-endpoint: unix:///run/containerd/containerd.sock
image-endpoint: unix:///run/containerd/containerd.sock
timeout: 10
debug: true

    2.2 containerd配置

    containerd config default > /etc/containerd/config.toml

# cat /etc/containerd/config.toml 

root = "/var/lib/containerd"
state = "/run/containerd"
oom_score = 0

[grpc]
  address = "/run/containerd/containerd.sock"
  uid = 0
  gid = 0

.........................................................................

    2.2.1 kata-container cri 插件配置，对于 1.5+版本

    plugins.cri.containerd.runtimes.runc: the runc, and it is the default runtime

    plugins.cri.containerd.runtimes.kata: The function in containerd (reference the document here) where the dot-connected string io.containerd.kata.v2 is translated to containerd-shim-kata-v2

[plugins.cri.containerd]no_pivot = false[plugins.cri.containerd.runtimes][plugins.cri.containerd.runtimes.runc]runtime_type = "io.containerd.runc.v1"[plugins.cri.containerd.runtimes.runc.options]NoPivotRoot = falseNOnewKeyring= falseShimCgroup = ""IoUid = 0IoGid = 0BinaryName = "runc"Root = ""CriuPath = ""SystemdCgroup = false[plugins.cri.containerd.runtimes.kata]runtime_type = "io.containerd.kata.v2"[plugins.cri.containerd.runtimes.katacli]runtime_type = "io.containerd.runc.v1"[plugins.cri.containerd.runtimes.katacli.options]NoPivotRoot = falseNOnewKeyring= falseShimCgroup = ""IoUid = 0IoGid = 0BinaryName = "/usr/bin/kata-runtime"Root = ""CriuPath = ""SystemdCgroup = false

    参考： https://github.com/kata-containers/documentation/blob/master/how-to/containerd-kata.md

     2.3 kubelet配置添加如下：

  --runtime-cgroups=/system.slice/containerd.service --container-runtime=remote --runtime-request-timeout=15m --container-runtime-endpoint=unix:///run/containerd/containerd.sock

     镜像下载的问题

ctr cri load xxx.tar

参考：

    https://github.com/kata-containers/runtime 

katacontainer 原理架构

Overview

    kata-container 和 runc是平级的。docker 管理容器生命周期的框架，而真正启动容器是 runc，可以换成 kata。kata-container可以当做 docker 的一个插件

参考：

    https://github.com/kata-containers/documentation/blob/master/design/architecture.md 

    https://yq.aliyun.com/articles/692117

   安装 https://github.com/kata-containers/documentation/blob/master/install/installing-with-kata-doc-to-script.md