Announcement
This blog is dedicated to building a "real home data center" at the lowest cost and exploring related technologies. In the process, I also hope to break the current situation where Internet blogs are dominated by "programmers' perspectives" and try to share more technical content from the perspectives of professional fields such as network engineering, operation and maintenance, and application delivery. Knowledge in these fields is often rarely discussed specifically (either too basic, or because the technical threshold of building a website makes it difficult for many technical personnel in related fields to share), and I hope to provide some different perspectives in these aspects and present some more in-depth technical content.
Hello blogger, after I used the proxy mode of warp, the direct access to https://ip.gs and the access through port 40000 all returned the address after warp NAT. Chatgpt's answer is that it may be in warp mode rather than proxy mode, but I read the situation written in your article, the remote access of vps will be directly interrupted after using warp mode. I tried to change it to warp mode and found that it was indeed the case. The connection was restored after changing it to proxy mode in vnc. Then why do two access methods return the same address after NAT? At the same time, I found that warp has opened a new wgcf port. When warp is closed, the local address of the captured packet is displayed as the address of this port, but when checking ip route, the default route is displayed as the real ip address, indicating that the wgcf port is not the default route. Can you please help me find out why these two situations occur?
Your result is abnormal and different from what I saw, so you may need to find the reason in your own environment. However, if you do see that the default route is correct and it is indeed in proxy mode, but the access effect is the warp address, then you should check whether you have used a global proxy to point to the warp proxy address but forgot to delete it; or you may try restarting. Generally, when encountering supernatural phenomena, if it were me, I would choose to restart and try again.
Hello, blogger. When I use my own laptop to install and test WARP on a cloud server with a debian 11 environment, there is no response for a long time after entering the warp-cli registration new registration command, and then I can no longer connect to the server. How can I solve this problem? I hope you can give me some advice!
Some time ago, warp was briefly revived for a few days but now seems to be down again. If you are accessing it directly from the domestic network, I think it will be very difficult.
I just found that it is OK after I operated it from the web terminal of the server console, and now warp can be used normally. But after I turned on the proxy mode of WARP according to your tutorial, I used tcpdump to capture packets. The captured packets contained QUIC, WireGuard, TLS1.3 and other traffic, but no MASQUE and HTTP traffic. I checked the WARP settings and found that the proxy protocol was MASQUE. Do you have any insights or opinions on this situation?
Isn't that right? Isn't MASQUE based on QUIC and TLS1.3? Before MASQUE, it was based on wireguard. I said in the article about warp using masque: "Most relevant to WARP is that QUIC can provide better performance on low-latency networks or high packet loss networks through packet merging and multiplexing. During the handshake process, QUIC packets in different contexts can be merged into the same UDP datagram, reducing the number of receptions and system interruptions; through multiplexing, QUIC can carry multiple HTTP sessions in the same UDP connection. WARP also benefits from the high confidentiality of QUIC, and TLS 1.3 is used in the protocol."
I see. What you mean is that the captured QUIC and TLS1.3 traffic is enough to show that the current WARP communication is through the MASQUE protocol, but the protocol of these traffic is displayed as QUIC or TLS1.3 in Wireshark, right?
I understand, thank you very much for your article and patience!
The blog is well written, but the JS is a bit dirty and doesn't work for anything. I don't know who it's disgusting.
Regarding the content of the blog post, I have opened up copying for the code-related parts. I don’t think it’s necessary to copy other descriptive content, and this is also one of the preventive measures to prevent the article from being copied and pasted to other places at will (of course, this behavior is not generous enough). However, in fact, this method is useless. The nature of HTML determines that the content that other people see with the browser is not under my control, so if you really want to copy, just install a browser plug-in. In addition, this function is actually a plug-in for WordPress. I installed it because of other related needs (I was considering closing this function for subscribing members to improve the happiness of subscribing members, but I was too lazy to do it later). Later, I got used to it and felt that there was nothing wrong with it. It can be regarded as a copyright reminder and a small means to let users come back often (although it is actually useless).
Hello, blogger, I recently debugged Debian 12 on a domestic soc. I found that after waking up the system from sleep mode in Debian 12, the Bluetooth node disappeared and the Bluetooth firmware needed to be loaded through commands. I want to start a script to execute this command after waking up from sleep mode. Do you have any ideas?
In the future, you can just ask chatgpt about this kind of question. Your description is quite clear. I asked for you. Chatgpt's answer is at the link:Unrivaled file sharingThe access password of the "Sleep Wake Up Enable Bluetooth Script.txt" file is the domain name of my blog.
Hello blogger, your blog loads very fast, can you introduce the server provider and optimization techniques?
I directly used my home broadband + idle equipment (mac-mini) at home to build the website, without using a cloud host. This is actually a configuration skill of cloudflare. I have actually written about it in my previous articles. Generally speaking, there are two ways: 1. Configure the traditional cloudflare cache rules 2. Use the advanced version of worker optimization to achieve a higher level of acceleration. However, the premise is that your domain name is hosted on cloudflare.
I have written a series of tutorials about cloudflare (there is a cloudflare learning map in the pinned article). The acceleration configuration is in the sixth part (cache rules) and the seventh part (worker optimization). However, I suggest that you read the series of tutorials from the beginning, because the point that a website needs to focus on is not just the access speed, security is also very important, so it is best to have a comprehensive understanding of cloudflare's solutions.
I wonder if the blogger has set up a group?＞﹏＜
You can join the TG group:https://t.me/tangwudiblog
trinitycore3.3.5 The client is on Windows, and the server is on a virtual machine. The client can search for the server but cannot log in. What is the reason?
Could you explain it more clearly? I don't know what client or server you are talking about?
Sorry, I'll try to be more detailed.https://blog.tangwudi.com/game/trinitycore180/, I am following your tutorial to build a World of Warcraft server. The environment is the same as the tutorial, which is a debian12 system under a virtual machine. The construction process is also the same as the tutorial. I did not encounter any problems and the server started normally. However, when I run wow.exe (wow-3.3.5) under the win11 system, I can log in with the created user and select the server (according to the tutorial, the server name should be Invincible Azeroth), but I cannot log in to the server. I don't know where the problem may be.?
Being able to log in means that the client and authserver process are communicating normally, and there is no problem with the auth library in the database. If you cannot log in to the server, if everything is normal before, it must be a problem with the communication between the worldserver process and the database. Is your database also on the same device as the authserver and worldserver, or is the database separated on another device?
It is on a device. Is there any error information I can check? I tried to find it, but couldn't find the cause.
I need to change bind-address = 127.0.0.1 in the /etc/mysql/mariadb.conf.d/50-server.cnf file in the Debian system to bind-address = local ip when connecting to the Debian database from win11. Otherwise, the database cannot be connected externally. However, after the change, authserver will fail to execute again. Is this information helpful?
It should help, but I don't have time to think about it now, I have to go out for dinner. Let me briefly talk about the startup process: the job of authserver is to read the information in the realmlist table in the auth database, and then let the client connect to the worldserver according to the value of the address item in the table. The key point is the value of the address item. If you want to change it, try changing it to your local IP address.
OK
Solved. The address of the realmlist table in my auth database should use the real IP address, but I used the loopback address 127.0.0.1. After changing it, it works. Thanks to the blogger for the tutorial!
Enjoy your own Azeroth :)