Summary: Good cost performance, the premise is to get the event price.
I just renewed a wave of fees on the Alibaba Cloud console, and it has been used for a year in a blink of an eye, so I want to talk about my experience in this year.
Configuration:
This is a lightweight application server that was prostituted during Double Eleven last year. The 2-core 2G configuration should belong to the younger brother. Although it has been developed for many years, this is the first cloud server of its own. The renewal price is 99/year, which can be renewed for three years, and the total can be used for 4 years (check the situation before moving when it expires, which is currently planned)
application:
The system chooses the built-in docker, which probably does these things:
1. Development environment
Mysql, minio, and redis are all for the convenience of their own development, and to avoid redeploying these environments in some cases.
For example, I have a surface 5 that has been with me for many years, but it is already difficult to run an idea. How can I bear to run some other software on it? Moreover, my own mobile office equipment is the only seedling, so deploying the supporting services to the cloud server relieves some pressure. As for the nginx running in docker, it is a waste, I will stop him after writing this article (on the role of summary).
Under low-frequency use, these applications are still relatively smooth, and it is obvious that 2 cores and 2G are completely sufficient.
Oh, by the way, remember to change the default port and password, don't ask why, it was scanned.
2、WordPress
I originally placed high hopes on wordpress and wanted to use it as my personal website. After actual use, the editor is not easy to use, the editor is very difficult to use, and the editor is extremely difficult to use. How can it be difficult to use as an article station editor? No Miss Li.
The other thing is that it feels a bit heavy, and the expansion is too much, but the user experience is a bit stuck. It may be that the small broken machine can't handle it a little bit.
3. FRP intranet penetration (really easy to use!)
Although the memory and CPU of the small broken machine are not enough, the 5M traffic is still very fragrant!
The original interface docking:
front end ==> run the background locally ==> Papapa reported an error ==> loudly notify the bi backend ==> the backend opens the frontend locally ==> follow the steps to debug ==> fix the bug = => Upload code ==> Front-end pull code ==> Front-end restart ==> Front-end test ==> Front-end found an error again...
Open the docking after penetration:
Front-end ==> Development encountered an error ==> Loud communication bi Knowing the backend ==> the backend reads an error, restarts after repairing => the frontend finds an error again...
Damn it, it saves so much workload for the frontend.
Even:
front end: the test server is down ah ah ah ah ah ah ah ah ah ah ah ah ah ah ah ah ah ah ah ah ah ah ah ah ah!
Backend: You even penetrate me.
Project Manager: There is no test machine to use spring cloud!
Backend: You even penetrate me.
Project Manager: There is no server to run the official project!
Backend: Go!
Originally, the penetration needs to be purchased, and it is rented one port at a time. Later, it was found that someone on the omnipotent github has open sourced how to build a frp service by itself (refer to the tutorial https://www.jianshu.com/p/1a30d9ec1775 )
. From then on, start up every day = => Penetration client activated!
recommend
To be honest, a 2h2c machine really can't do much, and even becoming a node of the cluster feels that it is not very suitable. But the cloud server is really delicious to use penetration! (Breaking sound)