After switching to WordPress, the choice of hosting services became wider. Over the years, in order to save money, I tried Digital Ocean, Linode, Vultr and other hosting services, and finally moved to BandwagonHost. I also installed WordPress on my own VPS at the beginning, and gradually switched to Docker compose.
Two weeks ago, I suddenly became interested in Kubernetes, so I used my algeria phone number data time to build a Kubernetes cluster on my Mac to learn, and gradually mastered the basic knowledge of Kubernetes. Then last week, I started to migrate Blogtang to Kubernetes on the host of BandwagonHost. Since the host of BandwagonHost is a single VPS with an annual fee of US$99, I can only build a Master Node, and the storage is only local. And because I run many services such as Gitea, Pypi, Jupyter Notebook (Tensorflow version), the resources are always stretched. I often see Pods being Evicted, which makes Blogtang services frequently interrupted. It's really miserable. Starting from Saturday, AWS was enabled, and I read a lot of AWS documents. I decided to use AWS EKS architecture, with two work nodes (two t2.medium EC2 instances), another EC2, mounted EBS hard disk, as NFS Server, and spent the whole day on Sunday to solve the problem of NFS connection time out, and finally solved it completely today (it turned out to be a problem with security group permission configuration). The Ingress solution uses Traefik, which integrates the automatic renewal service of Let's Encrypt's SSL certificate, which meets my requirements. MySQL currently uses the docker version of MariaDB, but it is planned to be migrated to AWS RDS later, and the database is also planned to be changed to Postgres, which I have been using for nearly a year.
Everything is running normally now, so I'll keep this as a record.