当前位置:网站首页>About my experience of "binary deployment kubernetes cluster"

About my experience of "binary deployment kubernetes cluster"

2022-06-11 18:54:00 Look, the future

experience

About this matter , Finally, it's over . If you start from installation , It took about twenty hours , It's a long job . If we do it from the beginning , That's really too long .

During this installation , I am right. kubernetes With a deeper understanding , I don't think it's too much to take this thing as a task .

Of course , Understanding is still relatively shallow , I don't agree that the information I found is shallow , I was too radical during this period of time , I haven't eaten well , Just arrive “ I see ”, Then the next part, It is impossible for me to retell .

But now I'm done deploying , I look back , There is still a month for me to slowly understand 、 Digestion 、 expansion .

I admit that this deployment has “ bet ” The ingredients are in it , not so bad , I bet right . I began to write the first article “ Environmental preparation ” When , In fact, I have only done docker Install over there , That is to say master The previous step of installation . That environmental installation has been on the hot list for several days , In case the later installation fails , That's the accident .

During this period, I also had the idea to retreat , In my master perform kubectl When it comes to testing , Say the certificate is wrong , Then I thought about looking at resources , It turned out that there was nothing , The solution level on Baidu should also be well understood by everyone , Otherwise I wouldn't have made one myself “ Solution ” special column , Just to be free from being controlled by others .

But I soon gave up the idea , I'm not too overwhelmed by this pressure , I can resist leg injuries for four years , Then cure .
But at that time, I really didn't know where the certificate went wrong , First, I don't know about these certificates , stay “ Preparation for the new term ” That article said , I'm going to sort out the certificate after I've written it all . So I stopped , Go through all the certificates . then , Restart deployment .

Here I want to remind myself , Want a full uninstall k8s Clustered . The reason why the deployment was restarted , This is because the previously deployed cluster will not be cleaned up , As a result, only two virtual machines can be reopened .

I suggest having the ability 、 Have an idea 、 A willing friend , You can experience , If you want to go Architecture , You have to do these things sooner or later .


No more nonsense , Let me sort out the blog in this process :

Blog finishing

Research stage

kubernetes framework
kubernetes Core components etcd
kubernetes API Access control : authentication
kubernetes API Access control : to grant authorization
kubernetes API Access control : Access control
kubernetes Core components APIServer
kubernetes Core components Controller Manager
kubernetes Core components Schedule
kubernetes Of components kubelet
kubernetes Of components kube-proxy

It's still short calico and coredns


Deployment

kubernetes Binary installation (v1.20.15)( One ) Environmental preparation
kubernetes Binary installation (v1.20.15)( Two ) Deploy etcd colony
kubernetes Binary installation (v1.20.15)( 3、 ... and ) Container runtime
kubernetes Binary installation (v1.20.15)( Four ) Deploy master
kubernetes Binary installation (v1.20.15)( 5、 ... and ) verification master Deploy
kubernetes Binary installation (v1.20.15)( 6、 ... and ) Deploy WorkNode node
kubernetes Binary installation (v1.20.15)( 7、 ... and ) Add a work node
kubernetes Binary installation (v1.20.15)( 8、 ... and ) Deploy The network plugin
kubernetes Binary installation (v1.20.15)( Nine ) finishing : Deploy several dashboards


Mid term supplementary research

kubernetes Certificate collection
see k8s Various versions Corresponding Version of each component
PKI/TLS Swiss Army knife cfssl
Kubernetes TLS bootstrapping


Expected research in the later stage

It will come out in a few days :

calico
coredns
 clear  k8s  colony 

Problem solution

Problem solving :[/usr/lib/systemd/system/etcd.service:6] Missing ‘=‘.
Problem solving :Failed to start etcd.service: Unit is masked.
[Failed to create pod sandbox: rpc error: code = Unknown desc = failed to set up sandbox container…

more :
Problem solving column


Let me rest for a few days …

原网站

版权声明
本文为[Look, the future]所创,转载请带上原文链接,感谢
https://yzsam.com/2022/162/202206111833348129.html