iperfInk8s

yaml

Like following:

# tcpprobe https://wiki.linuxfoundation.org/networking/tcpprobe
# use: apt install module-init-tools
# to install modprobe
apiVersion: extensions/v1beta1
kind: Deployment
metadata:
  annotations:
    deployment.kubernetes.io/revision: "1"
  generation: 1
  labels:
    run: ipref
  name: ipref
  namespace: default
spec:
  replicas: 12 # we have 10 nodes in a cluster hence 12 replicas
  selector:
    matchLabels:
      run: ipref
  template:
    metadata:
      labels:
        run: ipref
    spec:
      containers:
      - command:
        - sleep
        - "infinity"
        image: networkstatic/iperf3
        imagePullPolicy: Always
        name: ipref
        resources: {}
        securityContext:
          capabilities:
            add:
            - CAP_ALL
          privileged: true
        volumeMounts:
          - mountPath: /dev
            name: dev
          - mountPath: /lib/modules
            name: modules
      volumes:
      - name: dev
        hostPath:
          # directory location on host
          path: /dev
      - name: modules
        hostPath:
          # directory location on host
          path: /lib/modules

you should change the corresponding images and imagePullPolicy.

ntpoffline

Server

server side configuration:

# yum install -y ntpd
# vim /etc/ntpd.conf

The configuration file is listed as following:

driftfile /var/lib/ntp/drift
restrict default nomodify notrap nopeer noquery
restrict 127.0.0.1 
restrict ::1
restrict 192.168.0.0 mask 255.255.0.0 nomodify notrap
server 127.127.1.0  # local clock
fudge 127.127.1.0  stratum 10
includefile /etc/ntp/crypto/pw
keys /etc/ntp/keys
disable monitor

Disable the chronyd service thus the ntpd could acts properly:

# systemctl disable chronyd
# systemctl enable ntpd
# systemctl start ntpd
# systemctl disable firewalld

Client

Install via:

# yum install -y ntpd

Configuration file:

driftfile /var/lib/ntp/drift
restrict default nomodify notrap nopeer noquery
restrict 127.0.0.1 
restrict ::1
server 192.168.122.200
# 配置允许上游时间服务器主动修改本机的时间
restrict 192.168.122.200 nomodify notrap noquery
includefile /etc/ntp/crypto/pw
keys /etc/ntp/keys
disable monitor

Also disable the chronyd service and enable the ntpd service. The client will automatically sync with the server 192.168.122.200.

fabric8

什么是fabric8

fabric8是一个开源集成开发平台，为基于Kubernetes和Jenkins的微服务提供持续发布。可以认为它是一个对Java友好的开源微服务管理平台.

fabric8也可以被视为是一个微服务DevOps平台。Fabric8提供了一个完全集成的开源微服务平台，可在任何的Kubernetes和OpenShift环境中开箱即用。

参考:

https://jimmysong.io/posts/fabric8-introduction/

搭建过程(ArchLinux)

安装必要的包:

$ sudo pacman -S libvirt qemu dnsmasq ebtables

将自己的用户添加到kvm和libvirt用户组:

$ sudo usemod -a -G kvm,libvirt <username>

更新/etc/libvirt/qemu.conf中关于libvirt的配置:

$ sudo sed -r 's/group=".+"/group="kvm"/1' /etc/libvirt/qemu.conf > /etc/libvirt/qemu.conf

更新当前的session，以适配用户组改动:

$ newgrp libvirt

此外，我们需要在yaourt仓库中安装对应的包以使用dockermachine对于kvm的驱动:

$ sudo pacman -S docker-machine-kvm2 docker-machine
$ yaourt docker-machine-kvm

安装minishift:

$ yaourt minishift
$ minishift start --memory=7000 --cpus=4 --disk-size=50g

启动完毕后，可以检查对应的CPU/内存/磁盘信息等。

安装fabric8 on minishift(我用的是on-my-zsh):

$ echo 'export PATH=$PATH:~/.fabric8/bin' >> ~/.zshrc
$ source ~/.zshrc

配置GitHub Client ID/密码, 参考:

https://developer.github.com/apps/building-integrations/setting-up-and-registering-oauth-apps/registering-oauth-apps/

URL可以填写为:

http://keycloak-fabric8.{minishift ipv4 value}.nip.io/auth/realms/fabric8/broker/github/endpoint

homepage的URL可以填写为https://fabric8.io.

由上面得到的clientID和client secret可以被引入到环境变量中:

$ export GITHUB_OAUTH_CLIENT_ID=123
$ export GITHUB_OAUTH_CLIENT_SECRET=123abc

之后:

$ gofabric8 start --minishift --package=system  --namespace fabric8

经过漫长的等待(需要翻墙)， fabric8环境将就绪，用来登录的用户名/密码分别为"developer/developer”

fabric 8 playing

以system:admin登录，查看工作空间:

$ oc login -u system:admin -n default
Logged into "https://192.168.42.131:8443" as "system:admin" using existing credentials.

You have access to the following projects and can switch between them with 'oc project <projectname>':

  * default
    developer
    developer-che
    developer-jenkins
    developer-run
    developer-stage
    fabric8
    kube-public
    kube-system
    myproject
    openshift
    openshift-infra
    openshift-node

Using project "default".

可以看到，fabric8的namespaces已经被创建出来。

WorkingTipsForArchLinuxSpring

Installation

Install the following packages:

$ sudo pacman -S maven community/intellij-idea-community-edition
$ sudo pacman -S jdk8-openjdk jdk9-openjdk
$ sudo archlinux-java set java-9-openjdk
$ archlinux-java status

Since the intellij wants jdk8 or newer, you have to install newer jdk implementation.

Correct: the community edition didn’t have the spring boot support, use the ultimate edition:

$ yaourt intellij-idea-ultimate-edition

扫盲

什么是spring boot

spring boot 致力于简洁，让开发者写更少的配置，程序能够更快的运行和启动。它是下一代javaweb框架，并且它是spring cloud（微服务）的基础。

spring boot

Create new project:

Plugins:

Import project:

mvn aliyun configuration

In /opt/maven/conf.

KismaticGlusterfsTips

Configuration

The configuration files is listed as following:

# Worker nodes are the ones that will run your workloads on the cluster.
worker:
  expected_count: 1
  nodes:
  - host: "allinone"
    ip: "10.15.205.93"
    internalip: ""
    labels: {}

storage:
  expected_count: 3
  nodes: 
  - host: "gluster1"
    ip: "10.15.205.90"
    internalip: ""
    labels: {}
  - host: "gluster2"
    ip: "10.15.205.91"
    internalip: ""
    labels: {}
  - host: "gluster3"
    ip: "10.15.205.92"
    internalip: ""
    labels: {}

# A set of NFS volumes for use by on-cluster persistent workloads
nfs:
  nfs_volume: []

But it won’t startup, the reason is because Ubuntu have a bug of rpcbind, solved by:

# systemctl add-wants multi-user.target rpcbind.service
# systemctl enable rpcbind.service
# ufw disable

Then you should reboot all of the nodes.

verification

Create a new glusterfs volume and expose it in k8s as a PV use:

# kismatic volume add 10 storage01 -r 2 -d 1 -c="durable" -a *.*.*.*

New PVC:

kind: PersistentVolumeClaim
apiVersion: v1
metadata:
  name: my-app-frontend-claim
  annotations:
    volume.beta.kubernetes.io/storage-class: "durable"
spec:
  accessModes:
    - ReadWriteMany
  resources:
    requests:
      storage: 10Gi

Use pvc in a pod volume:

kind: Pod
apiVersion: v1
metadata:
  name: my-app-frontend
spec:
  containers:
    - name: my-app-frontend
      image: nginx
      volumeMounts:
      - mountPath: "/var/www/html"
        name: html
  volumes:
    - name: html
      persistentVolumeClaim:
        claimName: my-app-frontend-claim

Whenyou scale the pod out, each instance of the pod should have access to that directory.