In May of this year, I added a new broadband, so I bought a new router TL-R483G to aggregate the two broadband into one line for intranet use. The article at that time is here, but recently I added one of the broadband Split out because aggregation didn't work as it should and this router was the worst shopping flop I've had lately.

# review

There is nothing wrong with the router itself, it is very convenient to use, the operations are reasonable, and various functions can be used normally.

The reason why it is said to fail is my broadband, China Unicom broadband can apply for a public network IP, but it is not a fixed public network, it will jump randomly, but generally not less than 24 hours. But due to China Unicom's disgusting operation, the IP of one of my two broadband became more unstable, and the highest frequency even changed twice in half an hour.

I didn't notice this problem at first. Until one of my online services always reported an error occasionally, but it automatically recovered after a while. I checked carefully, the service itself is normal, the local network facilities are normal (I didn’t realize it was an IP problem when I first checked), and the cloud configuration is also normal, but the online rate is only about 80%, so something must be wrong.

# Troubleshoot

Last weekend, when I received an email alert, I was just bored, so I followed the clues to find out the cause of the error.

This service is an external network to access local files, and the path is also very simple. It is abstracted: external network client access domain name "DNS server resolves IP" and accesses the local server according to IP. DNS resolution comes with TP-link routers, it will be updated automatically and it is free.

At this point the domain name is no longer accessible. I immediately checked the domain name resolution and found that it was inconsistent with the IP obtained by my local server. Then I looked at the console of the router, and the IPs of the two broadband were indeed inconsistent with the IP obtained by domain name resolution. This is very strange, I looked at the WAN port situation again, TMD, one port turned out to be the IP obtained 2 minutes ago, of course the DNS resolution has not been updated, and then I looked for the reason.

The local router is not broken. In order to prevent misjudgment, I deliberately restarted the router and dialed again. Both ports have obtained new public network IPs. After about 5 minutes, the external network access has also been resolved. That means that DDNS is actually working normally, and on the router, the dial-up time of the two ports is the same.

When I thought that all this could be solved by restarting, reality gave me a ruthless iron fist. In the afternoon the alarm was received again. Looking at it again, the IP of that port was changed, and it took less than 1 minute to reconnect, so the analysis failed and the detection failed. But the other one will not change.

Finally, I called the broadband maintenance person, and the conclusion was that China Unicom’s home broadband has no commitment to the public network IP itself. , it is estimated that it belongs to the junk user group, so it is not as easy to use as the new one.

# solution

Take out the two broadband separately, so that it will not cause errors on the main route. I always change the IP, I use frp to transfer to the broadband that does not change the IP frequently, which can be regarded as an indirect solution to this problem. Broadband aggregation and load balancing are for commercial broadband, and home broadband is not easy to use.

ps: The router has passed the return period, and I am reluctant to sell second-hand, so I continue to use it. Next time when buying these more professional equipment, you must consider whether I need this function and whether this function can be used in my environment...