I Broke My Entire Internet to Fix a Node.js Port Conflict

I just solved a problem that has been haunting me for several days. Recently, my whole network started acting as if it were broken. For any websites I tried to visit, I always needed to refresh multiple times so that all the assets on the website could be fully loaded. Nearly all websites couldn't display anything before the first refresh; some even failed to display anything after 7-8 refreshes. Beyond websites, those applications that needed network connectivity failed to take action most of the time. The most common things I saw these days were different kinds of error messages.

If you've encountered the same problem and want to know directly how to solve it, you can click here to go straight to the solve part.

What I've Tried

I tried to use a mobile hotspot instead of the Wi-Fi I usually did, but it didn't work. So I started to think that there must be something wrong with my computer.

The first thing that came to my mind was DNS. I tried different DNS settings and went through so many DNS-related blogs or articles. Nothing works.

Cause I sometimes use a VPN, I began to consider if it was causing some problems. However, this hypothesis proved to be wrong very quickly, because the network still didn't work after I switched off the switch in the VPN program and removed any proxy settings it had set in the system settings panel.

At this moment, I felt utterly hopeless. I'd tried everything I knew and could, but nothing worked. Will I need to live with such a shitty network for the days ahead and never get the chance to experience fluent network again? I thought. My heart pressure was lifting every time I tried to visit a website or open a program that needed network.

The Clue?

At one point, it suddenly occurred to me that maybe I could check the VPN's logs to see if I could find any clues or tips which would help me figure out what the problem was. I selected an error log. It read something like:

...An operation on a socket could not be performed because the system lacked sufficient buffer space or because a queue was full.

Oh, god. What I'd done several days before suddenly came into my mind.

I worked on some web applications, which need port 3000 to run locally. Most of the time when I tried to start a server on port 3000, I would encounter the EACCES error I've been talking about in another blog. This EACCES error is because the Windows system had occupied port 3000 for temporary usage. I didn't want to free the port from the system every time I needed to start a server, so I searched on the internet to check if I could configure the range of ports the system could use.

Yeah, there were indeed such methods. At that time, I didn't realize that the system needs so many temporary ports to keep the network work well, so I changed the range to give it only around 1000 ports available (it used to be like 14000 or similar). Later, you know what happened. So this is the problem.

Solve

Since we've known what the problem came from, it's easy to solve it. Just reassign sufficient ports to the system:

// First check how many ports are available to the system.
netsh int ipv4 show dynamicport tcp
// Then reassign the range. Here I gave it about 14,000 ports
netsh int ipv4 set dynamicport tcp start=1024 num=15000

Everything works fine now.