Kubernetes Chapter 8: deploy NFS system with kubernetes to complete database persistence (kubernetes work practice class)

One 、 Preface

This paper belongs to Kubernetes Work practice , In the work , Need to complete kubernetes Persistence .

kubernetes Persistence includes four steps ：

(1) Automatically create a new directory ：pod take effect , Then it means to create two directories ( Add two properties ,selector and (1) tolerance )
(2) NFS Persistence ：MySQL and redis Be persistent , be nfs take effect
(3) If the namespace disappears, the specified directory of the namespace will be deleted automatically ：crontab Scheduled tasks and cleanup scripts take effect
(4) df -h /nfs Check the directory disk size , If not enough , Use soft links when /nfs The directory is soft linked to the directory with large disk , As a subdirectory under it

NFS A way of storing server files , There are other ways , such as cephs, Good blog references are as follows ：
k8s pv Of accessMode The way and persistentVolumeReclainPolicy Introduction of methods
https://www.jianshu.com/p/0fab432831b3
k8s Manage storage resources
http://t.zoukankan.com/zzzynx-p-11065612.html
K8S Chapter nine of the series （ Persistent storage ,emptyDir、hostPath、PV/PVC）
https://blog.csdn.net/qq_33261700/article/details/119549172

SpringBoot Source download ：k8s-demo

Two 、 Set up the machine NFS The server

1、 Set up well NFS Server on
2、NFS Label the node where the server is located label, Let me create the directory automatically Pod Assigned to yes NFS On this node of the server , Create a new directory on this node
3、 Write the cleaning script on the machine clean.sh And timing tasks crontab
4、df -h /nfs Check the directory disk size , If not enough , Use soft links when /nfs The directory is soft linked to the directory with large disk , As a subdirectory under it

2.1 Set up well NFS Server on

01 choice master Node as nfs Of server, So in master Node , Execute the following command

#  First step , start-up nfs and rpcbind
#  start-up nfs
systemctl status nfs ( Open if it exists  systmctl start nfs , But if it doesn't exist, install  yum -y install nfs-utils  and  systemctl start nfs)
systemctl status nfs （ Check the status again after startup , Successful status means successful startup ）
systemctl enable nfs  ( Set it to start automatically )

#  start-up rpcbind
systemctl restart rpcbind   ( restart )
systemctl enable rpcbind    ( Set to power on self starting )
systemctl status rpcbind  ( Check the status , Verify that the restart is successful )


#  The second step , establish nfs Directory and grant permissions  /nfs/data/  This directory is nfs ip That catalog 
#  establish nfs Catalog 
mkdir -p /nfs/data/
#  Grant authority 
chmod -R 777 /nfs/data

#  The third step , edit export File and save 
#  edit export file   about /nfs/data Catalog , Grant read and write permission 、 Root permissions 、 Synchronize data permissions 
vi /etc/exports
  /nfs/data *(rw,no_root_squash,sync)
  /nfs/data *(rw,no_root_squash,sync,no_subtree_check)  #  new edition nfs
#  Make configuration effective 
exportfs -r
#  Check in effect 
exportfs

#  Step four , verification rpcbind、nfs
#  see rpc Registration of services 
rpcinfo -p localhost
# showmount test  　　
# showmount The command is used to query NFS Information about the server  　　-e or --exports  Show NFS Output list of the server .
showmount -e master-ip  
showmount -e 192.168.100.151

02 all node Install the client on ps -ef|grep nfs

#  start-up nfs
systemctl status nfs ( Open if it exists  systmctl start nfs , But if it doesn't exist, install  yum -y install nfs-utils  and  systemctl start nfs)
systemctl status nfs （ Check the status again after startup , Successful status means successful startup ）
systemctl enable nfs  ( Set it to start automatically )

#  start-up rpcbind
systemctl restart rpcbind   ( restart )
systemctl enable rpcbind    ( Set to power on self starting )
systemctl status rpcbind  ( Check the status , Verify that the restart is successful )

2.2 NFS Label the node where the server is located label

#  tagging 
kubectl label nodes w1 nfs=server
#  test 
kubectl describe node w1 | grep nfs

NFS Label the node where the server is located label, Then create a new directory automatically Pod.yaml Written in the document nodeSelector, You can automatically create a new directory Pod Dispatch (scheduler) To NFS Server node , Finish creating new directory ,

2.3 Clean up scripts and scheduled tasks

Clean up scripts ：shell Write a language clean.sh Script , The logic is “ For variables /nfs/data The following table of contents , If this namespace does not exist , Just delete this directory ”.

Timing task
systemctl status crond ( Start when you have , No installation )
vi /etc/crontab
systemctl start crond
systemctl enable crond

2.4 Soft link operation

Find one with disk df -h Check the remaining capacity of each disk
Create a soft connection ln -s Actual catalog Shortcut
View soft links df -h /nfs

df -h /nfs Check the remaining capacity of the directory
du -h /nfs Check the disk capacity

There is no need to hang the tray , There is still left 8.4G, Show a hanging tray

3、 ... and 、 springboot Connect k8s Cluster and write

Official website ：http://fabric8.io/
github： https://github.com/fabric8io/kubernetes-client
fabric8： Using document https://github.com/fabric8io/kubernetes-client Below readme.txt

Be careful , In general, we use fabric8 Below kubernetes-client

Okay , One springboot The project is set up .

If you will fabric8 Integrated into the springboot in , You can first refer to others , as follows ：

Chinese references ,http://t.zoukankan.com/larrydpk-p-14928606.html
Chinese references ( Through local or network yaml File to create resources to k8s On the cluster )：https://www.jianshu.com/p/48e0896b1608
Chinese references ( take token Put it in a file or data table )：https://blog.csdn.net/weixin_43962314/article/details/110094808

Okay , I began to write by myself .

The successful running , as follows ：

Four 、 The end of the

finish!!

Learn technology well, Day Day Up !