Lxcfs资源视图隔离

1. 资源视图隔离

容器中的执行topfree等命令展示出来的CPU,内存等信息是从/proc目录中的相关文件里读取出来的。而容器并没有对/proc/sys等文件系统做隔离,因此容器中读取出来的CPU和内存的信息是宿主机的信息,与容器实际分配和限制的资源量不同。

/proc/cpuinfo
/proc/diskstats
/proc/meminfo
/proc/stat
/proc/swaps
/proc/uptime

为了实现让容器内部的资源视图更像虚拟机,使得应用程序可以拿到真实的CPU和内存信息,就需要通过文件挂载的方式将cgroup的真实的容器资源信息挂载到容器内/proc下的文件,使得容器内执行top、free等命令时可以拿到真实的CPU和内存信息。

2. Lxcfs简介

lxcfs是一个FUSE文件系统,使得Linux容器的文件系统更像虚拟机。lxcfs是一个常驻进程运行在宿主机上,从而来自动维护宿主机cgroup中容器的真实资源信息与容器内/proc下文件的映射关系。

lxcfs的命令信息如下:

#/usr/local/bin/lxcfs -h
Usage:

lxcfs [-f|-d] -u -l -n [-p pidfile] mountpoint
  -f running foreground by default; -d enable debug output
  -l use loadavg
  -u no swap
  Default pidfile is /run/lxcfs.pid
lxcfs -h

lxcfs的源码:https://github.com/lxc/lxcfs

3. Lxcfs原理

lxcfs实现的基本原理是通过文件挂载的方式,把cgroup中容器相关的信息读取出来,存储到lxcfs相关的目录下,并将相关目录映射到容器内的/proc目录下,从而使得容器内执行top,free等命令时拿到的/proc下的数据是真实的cgroup分配给容器的CPU和内存数据。

原理图

lxcfs

映射目录

类别 容器内目录 宿主机lxcfs目录
cpu /proc/cpuinfo /var/lib/lxcfs/{container_id}/proc/cpuinfo
内存 /proc/meminfo /var/lib/lxcfs/{container_id}/proc/meminfo
/proc/diskstats /var/lib/lxcfs/{container_id}/proc/diskstats
/proc/stat /var/lib/lxcfs/{container_id}/proc/stat
/proc/swaps /var/lib/lxcfs/{container_id}/proc/swaps
/proc/uptime /var/lib/lxcfs/{container_id}/proc/uptime
/proc/loadavg /var/lib/lxcfs/{container_id}/proc/loadavg
/sys/devices/system/cpu/online /var/lib/lxcfs/{container_id}/sys/devices/system/cpu/online

4. 使用方式

4.1. 安装lxcfs

环境准备

yum install -y fuse fuse-lib fuse-devel

源码编译安装

git clone git://github.com/lxc/lxcfs
cd lxcfs
./bootstrap.sh
./configure
make
make install

或者通过rpm包安装

wget https://copr-be.cloud.fedoraproject.org/results/ganto/lxc3/epel-7-x86_64/01041891-lxcfs/lxcfs-3.1.2-0.2.el7.x86_64.rpm;
rpm -ivh lxcfs-3.1.2-0.2.el7.x86_64.rpm --force --nodeps

查看是否安装成功

lxcfs -h

4.2. 运行lxcfs

运行lxcfs主要执行两条命令。

sudo mkdir -p /var/lib/lxcfs
sudo lxcfs /var/lib/lxcfs

可以通过systemd运行。

lxcfs.service文件:

cat > /usr/lib/systemd/system/lxcfs.service <<EOF
[Unit]
Description=lxcfs

[Service]
ExecStart=/usr/bin/lxcfs -f /var/lib/lxcfs
Restart=on-failure
#ExecReload=/bin/kill -s SIGHUP $MAINPID

[Install]
WantedBy=multi-user.target
EOF

运行命令

systemctl daemon-reload && systemctl enable lxcfs && systemctl start lxcfs && systemctl status lxcfs 

4.3. 挂载容器内/proc下的文件目录

docker run -it --rm -m 256m  --cpus 2  \
      -v /var/lib/lxcfs/proc/cpuinfo:/proc/cpuinfo:rw \
      -v /var/lib/lxcfs/proc/diskstats:/proc/diskstats:rw \
      -v /var/lib/lxcfs/proc/meminfo:/proc/meminfo:rw \
      -v /var/lib/lxcfs/proc/stat:/proc/stat:rw \
      -v /var/lib/lxcfs/proc/swaps:/proc/swaps:rw \
      -v /var/lib/lxcfs/proc/uptime:/proc/uptime:rw \
      nginx:latest /bin/sh

4.4. 验证容器内CPU和内存

# cpu
grep -c processor /proc/cpuinfo
cat /proc/cpuinfo

# memory
free -g
cat /proc/meminfo

5. 使用k8s集群部署

使用k8s集群部署与systemd部署方式同理,需要解决2个问题:

  1. 在每个node节点上部署lxcfs常驻进程,lxcfs需要通过镜像来运行,可以通过daemonset来部署。
  2. 实现将lxcfs维护的目录自动挂载到pod内的/proc目录。

具体可参考:https://github.com/denverdino/lxcfs-admission-webhook

5.1. lxcfs-image

Dockerfile

FROM centos:7 as build
RUN yum -y update
RUN yum -y install fuse-devel pam-devel wget install gcc automake autoconf libtool make
ENV LXCFS_VERSION 3.1.2
RUN wget https://linuxcontainers.org/downloads/lxcfs/lxcfs-$LXCFS_VERSION.tar.gz && \
	mkdir /lxcfs && tar xzvf lxcfs-$LXCFS_VERSION.tar.gz -C /lxcfs  --strip-components=1 && \
	cd /lxcfs && ./configure && make

FROM centos:7
STOPSIGNAL SIGINT
COPY --from=build /lxcfs/lxcfs /usr/local/bin/lxcfs
COPY --from=build /lxcfs/.libs/liblxcfs.so /usr/local/lib/lxcfs/liblxcfs.so
COPY --from=build /lxcfs/lxcfs /lxcfs/lxcfs
COPY --from=build /lxcfs/.libs/liblxcfs.so /lxcfs/liblxcfs.so
COPY --from=build /usr/lib64/libfuse.so.2.9.2 /usr/lib64/libfuse.so.2.9.2
COPY --from=build /usr/lib64/libulockmgr.so.1.0.1 /usr/lib64/libulockmgr.so.1.0.1
RUN ln -s /usr/lib64/libfuse.so.2.9.2 /usr/lib64/libfuse.so.2 && \
    ln -s /usr/lib64/libulockmgr.so.1.0.1 /usr/lib64/libulockmgr.so.1
COPY start.sh /
CMD ["/start.sh"]

star.sh

#!/bin/bash

# Cleanup
nsenter -m/proc/1/ns/mnt fusermount -u /var/lib/lxcfs 2> /dev/null || true
nsenter -m/proc/1/ns/mnt [ -L /etc/mtab ] || \
        sed -i "/^lxcfs \/var\/lib\/lxcfs fuse.lxcfs/d" /etc/mtab

# Prepare
mkdir -p /usr/local/lib/lxcfs /var/lib/lxcfs

# Update lxcfs
cp -f /lxcfs/lxcfs /usr/local/bin/lxcfs
cp -f /lxcfs/liblxcfs.so /usr/local/lib/lxcfs/liblxcfs.so


# Mount
exec nsenter -m/proc/1/ns/mnt /usr/local/bin/lxcfs /var/lib/lxcfs/

5.2. daemonset

lxcfs-daemonset.yaml

apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: lxcfs
  labels:
    app: lxcfs
spec:
  selector:
    matchLabels:
      app: lxcfs
  template:
    metadata:
      labels:
        app: lxcfs
    spec:
      hostPID: true
      tolerations:
      - key: node-role.kubernetes.io/master
        effect: NoSchedule
      containers:
      - name: lxcfs
        image: registry.cn-hangzhou.aliyuncs.com/denverdino/lxcfs:3.1.2
        imagePullPolicy: Always
        securityContext:
          privileged: true
        volumeMounts:
        - name: cgroup
          mountPath: /sys/fs/cgroup
        - name: lxcfs
          mountPath: /var/lib/lxcfs
          mountPropagation: Bidirectional
        - name: usr-local
          mountPath: /usr/local
      volumes:
      - name: cgroup
        hostPath:
          path: /sys/fs/cgroup
      - name: usr-local
        hostPath:
          path: /usr/local
      - name: lxcfs
        hostPath:
          path: /var/lib/lxcfs
          type: DirectoryOrCreate

5.3. lxcfs-admission-webhook

lxcfs-admission-webhook实现了一个动态的准入webhook,更准确的讲是实现了一个修改性质的webhook,即监听pod的创建,然后对pod执行patch的操作,从而将lxcfs与容器内的目录映射关系植入到pod创建的yaml中从而实现自动挂载。

deployment

apiVersion: apps/v1
kind: Deployment
metadata:
  name: lxcfs-admission-webhook-deployment
  labels:
    app: lxcfs-admission-webhook
spec:
  replicas: 1
  selector:
    matchLabels:
      app: lxcfs-admission-webhook
  template:
    metadata:
      labels:
        app: lxcfs-admission-webhook
    spec:
      containers:
        - name: lxcfs-admission-webhook
          image: registry.cn-hangzhou.aliyuncs.com/denverdino/lxcfs-admission-webhook:v1
          imagePullPolicy: IfNotPresent
          args:
            - -tlsCertFile=/etc/webhook/certs/cert.pem
            - -tlsKeyFile=/etc/webhook/certs/key.pem
            - -alsologtostderr
            - -v=4
            - 2>&1
          volumeMounts:
            - name: webhook-certs
              mountPath: /etc/webhook/certs
              readOnly: true
      volumes:
        - name: webhook-certs
          secret:
            secretName: lxcfs-admission-webhook-certs

具体部署参考:install.sh

#!/bin/bash

./deployment/webhook-create-signed-cert.sh
kubectl get secret lxcfs-admission-webhook-certs

kubectl create -f deployment/deployment.yaml
kubectl create -f deployment/service.yaml
cat ./deployment/mutatingwebhook.yaml | ./deployment/webhook-patch-ca-bundle.sh > ./deployment/mutatingwebhook-ca-bundle.yaml
kubectl create -f deployment/mutatingwebhook-ca-bundle.yaml

执行命令

/deployment/install.sh

参考:


最后修改 December 25, 2022: deploy by blog source (a162b04)