I could have happily spent the National Day --, but since the server was torn down yesterday afternoon and the car overturned, I had to make up for it. Until now, it has almost been repaired~~~

The reason why it is called almost repaired is that some data cannot be recovered. This time, it is really a long memory.
The reason is that yesterday afternoon I tried to upload the attachment of the blog to Qiniu Cloud through the plug-in, which is very fast. I also tried to change the directory to another space for easy access next time,
I have always used the bt panel, and I think it is convenient. However, after I changed the space, I could not even connect to the seven cow cloud. It is still not possible to uninstall and reinstall the bt plug-in. It is always abnormal. I tried to repair the bt panel again. After the repair, something magical happened, and the panel could not be opened directly!!!! In an instant, I was stunned, so I directly checked the server on the xshell. The bt panel process was dead and dead. I reported a python related program, and I remembered that I upgraded the server python environment around the 20th of last month, and wrote a special article -- the upgrade was successful, but the server was not restarted!!! Maybe this panel repair led to a conflict after restarting the panel. I searched for data for a long time, but I still couldn't solve it last night -- many online data said that a plug-in in python collided with bt, which was hopeless. When upgrading python, I didn't notice the backup, so I was embarrassed I don't know much about it. It's hard to re install the environment. My last way to restart the server -- I expected that it might not start before restarting, so I backed up some of the data, but not all of the data, so I lost some of it. It's true that after restarting, xshell couldn't connect. I slept too late yesterday, and the website continued to gg all night, Including this blog, which was recently migrated. I'm worried about the data loss, but fortunately, the blog data has not been lost. Fortunately, until today, the server has been reinstalled and all the environments and deployments have been completed. At present, the blog, note stack and movies have been restored. He is not panicking for the time being. Go to have a rest first hahahaha.

However, this thing really taught me a lesson. I've never done such a big thing before. This time, I lost some data in the note stack, which is only used for long-term memory.
The following must be done:
1. Be sure to back up the backup before upgrading and major operations. You can't predict what will happen after your operations.
2. After all operations are completed, the server must be restarted to check whether there are exceptions. Many times, restarting will cause many problems.
3. Do a scheduled backup to prevent data loss, which can be backed up to FTP or object storage space.
4. Do permission management on the server, and do not log on to the root user at random to avoid deleting wrong data.