当前位置：网站首页>Apache2 ha experiment with raspberry pie

Apache2 ha experiment with raspberry pie

2022-07-24 13:46:00 【Ki1381】

I heard that raspberry pie is also a financial product , I can't help but fall into meditation after silently closing the website of a treasure , Then I dug out the pile I bought before 3B. Come back , Although the book is full , But I'm not happy . According to Guo Degang , I'm so happy that this pile of raspberry pie can't cure me . Or choose two to make one HA Your experiment , It may be a little happy to make it .

So try this scenario ：
A website assumption is very important , When one server hangs up, another one can be on top

The following is the basic information of the experimental equipment ：
node 1：hostname ki01;IP：192.168.1.81
node 2：hostname ki02;IP：192.168.1.82
Cluster these two nodes , Virtual out a new address 192.168.1.83. Visit the website and use this address .

How to set fixed IP, How to set hostname Just skip it .

【 The following operations should be performed on both nodes 】

1、 modify /etc/hosts, Forced to join ki01 and ki02 Corresponding IP

192.168.1.81    ki01
192.168.1.82    ki02

2、 Installation required HA Service and make simple initial configuration ：

sudo apt install pacemaker pcs resource-agents
sudo systemctl start pcsd.service 
sudo systemctl enable pcsd.service
sudo passwd hacluster

  Set up hacluster And remember .

【 The following operations only need to be operated on any node 】

3、 Form a cluster 

#  Forced clean up 
sudo pcs cluster destroy
#  Register cluster nodes and form a cluster 
sudo pcs --force host auth ki01 ki02 -u hacluster -p  The password you just set  
sudo pcs --force cluster setup kicluster ki01 ki02

4、 Start the cluster and set it to start automatically after power on ：

sudo pcs cluster start –-all
sudo pcs cluster enable --all

5、 Ban stonith And arbitration （ For this experiment, none of these is necessary ）

sudo pcs property set stonith-enabled=false
sudo pcs property set no-quorum-policy=ignore

【 The following operations should be performed on both nodes 】

6、 install Apache2, And set to prohibit automatic startup ：

sudo apt install -y apache2
sudo systemctl stop apache2
sudo systemctl disable apache2

7、 edit /var/www/html/a.html, In order to observe the effect （ This is to ki01 For example .ki02 Just change KI02）：

<html>
 <body>THIS IS KI01.</body>
</html>

8、 see  /etc/apache2/mods-available/status.conf, Confirm the following statement ：

<Location /server-status>
    SetHandler server-status
    Require local
 </Location>

【 The following operations only need to be operated on any node 】

9、 Configure cluster resources ：WebSite（ Corresponding Apache2） And the virtual IP

#  establish  WebSite  Cluster resources 
sudo pcs --force resource create WebSite ocf:heartbeat:apache  \
      configfile=/etc/apache2/apache2.conf \
      statusurl="http://localhost/server-status" \
      op monitor interval=1min

#  establish   fictitious IP  Cluster resources 
sudo pcs --force resource create ClusterIP ocf:heartbeat:IPaddr2 ip=192.168.1.83

10、 Set up WebSite And the virtual IP Dependency of 、 Startup sequence 、 Node weight, etc . Normally, it should WebSite And the virtual IP It is on a node ：

sudo pcs constraint colocation add WebSite with ClusterIP INFINITY
sudo pcs constraint order ClusterIP then WebSite
sudo pcs constraint location WebSite prefers ki01=50

11、 It can be used at any time  sudo pcs status View the status of the cluster . Focus on nodes online/offline state 、 The state of the resource .

  At this time, you can use the browser to access http://192.168.1.83/a.html. At the moment, resources are ki01 On , Then it will show THIS IS KI01.

12、 Manual switching test （ The following statement is for the current resource in ki01 Upper ）：

sudo pcs resource move WebSite ki02

Use sudo pcs status command , And browser access 192.168.1.83 See if it has been cut .

13、 Automatic switching test ：

Use sudo pcs status Command or access 192.168.1.83 Determine which node the current resource is on . Then close the node . Normally, you should switch to the surviving node soon . Continue to use on surviving nodes sudo pcs status command , And browser access 192.168.1.83 Confirm whether it works normally .

When the closed node is reopened and online , Because the priority is set ki01=50（ The larger the number, the higher the priority ）, Resources will still be switched back ki01：

 

Reference material ：
https://clusterlabs.org/quickstart-redhat.html
pcs(8) — pcs — Debian unstable — Debian Manpages

原网站

版权声明
本文为[Ki1381]所创，转载请带上原文链接，感谢
https://yzsam.com/2022/205/202207241345121956.html

边栏推荐

猜你喜欢

随机推荐