Scribbles for my bad memory

Cause:

This morning, I change to the /backup directory and found I cannot write in it. So I rapidly typed sudo chown -R david:david .. then press Enter. You know what happened! All the files in / directory were changing the owner to me! I realized this by seeing an error message like "You cannot change the owner of xxx file to david".

Damn, how silly I was! Unfortunately, that was not the end of my bad luck. When I try to revert it by typing sudo chown -R root:root /, I got another error message: /usr/bin/sudo must be owned by uid 0 and have the setuid bit set. Then I tried su - root but it seems that I haven't set a root password.

How can I recover my pop! OS?

1. reboot it into single-user mode, edit the boot menu, add systemd.unit=rescue.target at the end of the boot line.
2. In the boot console, chown -R root:root /usr; chmod 4755 /usr/bin/{sudo,dpkg,pkexec,crontab}; reboot

Rather easy, right? It did cost me over 10mins! PS. I met many issues later after recovering from the accident. One of them is that the crontab command did not work as usual. I have to run the instruction below to fix it.

sudo apt reinstall cron

Linux is also fragile, please don't be such silly thing next time. I told myself. PS: This time I also set a root password as well, so next time I could run su - root directly instead of going to the single user mode.

In short:

Change the service file, and add two lines after [Service] line,

[Service]
LimitNOFILE=65535
LimitNOFILESoft=65535

If you want to know more, read the Reference

Just record what I made Nginx working with a project located in someone's HOME directory on a Linux machine with SELinux on.

1. Nginx seems working normally, but it actually Didn't listen to a non-80 port at all. If it reports ValueError: Port tcp/8081 already defined, replace -a with -m.

semanage port -a -t http_port_t -p tcp 8081

Reference1 Reference2

2. telnet localhost 8081 works, but telnet 192.168.168.168 8081 from another machine not working!

firewall-cmd _permanent add-port=8081/tcp

firewall-cmd --reload

3. Everything seems working good, but when you visit your site, Nginx just gives you a 403! You should ensure Nginx can access your project directory, everyone knows that, but is not enough when SELinux is on.

setsebool -P httpd_enable_homedirs 1

setenforce 0
systemctl restart nginx
systemctl daemon-reload

Reference

I had written the following code in a project

   ...
   return s && s.value || null

It works well for some days until a colleague did some code refactoring. In the beginning, s.value is a string value, and an empty string is not a valid value, so the code works well. After the refactoring, s.value became an integer, and this time 0 is a valid value. So you can imagine when s.value === 0 the code above will return null instead of 0. It leads to a bug!

Therefore please use the"A || B" expression with caution!