Serious and annoying packet loss

The company's sales office is connected to the central data center through a single-mode optical fiber cable. The core switch of the Central Data Center is the Passport 8606 switch of beidian. The access layer switch of the sales office is the BayStack425 switch of beidian power, and the sales system server is HP370, place it in the central data center and connect it to the 8606 switch. The client computer accesses the house sales system through IE browser to perform the house sales operation.
When all computers in the sales office use the sales system in one day, the homepage can be opened and users and passwords can be entered to log on to the system. However, the webpage is slow when further operations are performed, or the "this page cannot be displayed" prompt appears, and the page is displayed slowly after several refreshes, seriously affecting normal operations.
I first suspect that there is a problem with the server. After the server is restarted, the test still fails. Then, the computer in the central data center and the sales office can ping each other and test the connection without packet loss. Then, I directly access and operate the sales system on the workstation in the Central Data Center. Everything is normal, it indicates that the network of the sales server and the central data center is normal. Is there a network problem at the sales office? Try to restart the switch at the sales office and test that the fault still exists.
Are all computers in the sales office infected with viruses? I brought another laptop to the sales office for testing, and the same fault still occurred. When trying to ping the 65500 big package to the sales server, I finally found the problem-ping the big package from time to time cannot be connected, packet loss is serious-the loss rate is 40%. Ping 65500 big packages from the sales server to the sales office computer, packet loss is equally serious, but the sales office computer ping each other 65500 big package did not appear packet loss, Average value is 13 ms, it indicates that the sales office switch is no problem. At this point, I determined that the problem was caused by a fiber jumper or a cable connecting the two office buildings. First, check the optical fiber jumper on the switch at the sales office. It is found that there is a fiber jumper connector that is loose with the plug-in of the optical fiber module and it does not feel tight. The optical fiber connector is fastened with one click by hand. Then, the fault disappears.
Through this fault solution, we hope that when you encounter a network fault, you will not simply be able to ping the network to determine whether the network is normal, but should also send the largest packet test, to reach the final conclusion.
The cause of this fault is that optical signals are used for optical fiber transmission data, while Optical Fiber jumper connectors are not fully in contact with the optical fiber module, resulting in poor optical signal reception, and thus attenuation is particularly harmful, packet Loss is serious, which leads to problems in the client's access to the house sales system.

Related Articles]

• Network Management Practice: exploring network connection packet loss at different layers

• What is packet loss? Why?