Top 10 common LAN errors and Solutions

Network faults are inevitable. It is important to quickly isolate and eliminate faults. Network Maintenance personnel should be equipped with appropriate tools and knowledge to locate and solve problems in a timely and effective manner. This article discusses ten headaches that network technicians and engineers often encounter. Analyze the problem first, and then discuss the tools used to help find and isolate faults. It aims to serve as an example.

Problem 1: arrest an abuse of the Physical Layer

-User complaint-login failure or intermittent interruption
-Symptom: Only one workstation is affected, and the station is normal.

-Find the problem:

The Cable Terminal Link is too long. Users move, bend, connect, and disconnect network terminals every day (especially laptops ). A cheap cable tester can check this problem. Before you can say it is a cable problem, it is usually not a cable problem first. This requires the use of network testing tools to prove whether the region network segment is in good health, whether the hub port function is normal, and whether the nic and its Driver work well. Finally, check whether the network interconnection device is faulty. Only at this time will you know where to test the cable to isolate faulty terminals, disconnected connectors, ground circuits, and other issues. In the copper Axis Network, find the broken head between the coaxial "T-head" and the network card, and continuously run the instrument graph test function in the u tp network, continuously twist the terminal to identify poor contact or short circuit.

-Solve the problem:

Change workstation connection cables, replace damaged connectors, or stop using horizontal cables if necessary. Check the processed cables thoroughly again.

Problem 2: Fast Ethernet is slower than the Ethernet before the upgrade

-User complaint-10 Mbps Ethernet speed is slow. After upgrading to Mbps, you cannot connect to the Internet.
-Symptom-new workstation or upgraded workstation connection is affected.

-Find the problem:

For UTP5 unshielded twisted pair wires, use a cable tester to test whether it meets the EIA/TIA-TSB67 standard for the corresponding Mbps transmission rate. Some cable links that work normally over 10 Mbps Ethernet cannot work normally over Mbps Ethernet due to the high degree of near-end crosstalk. Signal coupling to adjacent line pairs causes high-frequency signal transmission failure. Separating UTP line pairs can even paralyze the network when traffic reaches a certain level.

-Solve the problem:

Replace or stop links that cannot be connected. After the fault is cleared, all links that have been tested and disposed of will be replaced.

 

Problem 3: Invalid Watchdog

-User complaint-unable to connect to the Netware remote server of the Novell Network
-Symptom-affects all workstations connected to servers in this segment

-Find the problem:

Check the health status of the CIDR block. Observe the usage, collision, error frame, and broadcast frame levels. At this time, we can observe a large number of collision or error frames. This collision or error frame may cause a delay between the "Hello" frame between the client and the server. Repeat this test on the remote network segment of the server specified by a certain device. If both CIDR blocks are blank, use the network test tool to query the SNMP agent in the vro and check the route table between the two CIDR blocks. Make sure the path does not have a congestion problem. To effectively use resources, if the client does not respond to the "Hello" frame within the time allowed by the watchdog, the Novell Server will deregister the client.

-Solve the problem:

In this method, the parameter delay of the watchdog is changed, and the value of this parameter in the Novell Server is increased. If the situation improves, you can correct the parameter of the frame timer with the watchdog delay to determine the cause of Frame loss.

Question 4: Chrysanthemum Effect

-User complaint-network service slowdown
-Symptom-affects all users of an Ethernet segment. The Network Service (Access Server) slows down, but it has nothing to do with a connected server.

-Find the problem:

Connect the network tester to the Hub during peak hours to run the network health test. A large number of error frames are observed. In the 10BASE-2 network, the error frame type is classified as "latency Collision ". In the 10BASE-T network, frame errors are classified as typical frame read sequence errors (FCS errors). As network users increase, new hubs usually have an effect on the original hubs.

-Solve the problem:

Count the number of reconnections between any two network devices. In 10 Mbps Ethernet, there should be no more than four, and in NBPs fast Ethernet, there should be no more than two. If the number of frames exceeds the upper limit, the network topology is modified according to the Ethernet standard, and the error frame is displayed again.

 

Question 5: retribution of IP network managers

-User complaint-a workstation gap error occurs. You cannot log on to or access the remote server.
-Symptom: Only one workstation is affected. This problem usually occurs after the workstation is added, moved, or changed.

-Find the problem:

First, eliminate potential account and Security Settings issues. Allow the user to log in and try to repeat the operation from another workstation. If the connection is successful, you can use the network tester to check the link between the client and the server. Then test the NIC to check the functions and status of the hardware and driver. If there is no problem, check the IP settings to determine whether the user's IP address is valid or whether there are duplicate IP addresses. Determine whether the subnet mask and default router address of this segment are correct, and whether the DNS server address (if any) is dedicated. If the IP address of a new workstation is the same as the IP address of the printing server, the printing work of each user is directed to the new client.

-Solve the problem:

Change to the appropriate settings, and then restart the workstation.

Question 6: Optical Fiber illusion

-User complaints-the entire network slows down
-Symptom-affects all workstations connected to network segments

-Find the problem:

During normal network usage, the network tester is connected to the hub to generate traffic at both ends of the optical fiber connection and the health status of each side. When the optical fiber is disconnected or connected at a certain time, a large number of error frames will be observed whenever the optical fiber is connected to the network segment. Poor fiber link connections generate a large number of junk frames in their affiliated CIDR blocks, forcing the workstation to resend frames and slowing down the network speed.

-Solve the problem:

Clean or reinstall the optical fiber connector, reset all connectors, and check the network health again. At this time, there should be only a small number of error frames.

 

Question 7: Strange! Is the cable hit at night?

-User complaint-unable to connect to the Internet or login
-Symptom-only one workstation can access the Internet on the network segment

-Search for Problems

Disconnect the connection cable of the workstation and connect it to the network tester. Note that no connection signal is detected from the hub. Run the cable test to check whether the cables are open from the Office to the wiring room. If the length is normal, set the tool to the location of the audio generator and use an audio detector from the wiring room to find the cable corresponding to the office. Normally, when there is a change in the wiring room, the cable leads will be pushed and collided, causing the connector to fall off or loosen.

-Solve the problem:

Re-plug the cable connector between the workstation and the hub to check whether the connection signal indicator is bright and whether the user can access the Internet.

Question 8: is the patient's heart still jumping?

-User complaint-the network access speed is slow during the busy network billing period.
-Symptom-All workstations in some CIDR blocks do not respond to servers in either the CIDR block or other CIDR blocks.

-Find the problem:

Connect the network tester to this network segment during peak hours. Run the network health test for this CIDR block. Pay attention to the network utilization, collision, error frames, and broadcast frames. At this time, the average collision rate may be abnormal, up to 50% of the observed traffic may be collision frame fragments. This is an error in sending and receiving frame signals-an SQL (Signal quality error. The "Heartbeat" signal is inadvertently stuck in the "Bo" position.

-Solve the problem:

Locate each signal sender in the network. Check whether the SQE switch is in the off position. This signal was used in the early 10BASE-5 Ethernet, and the new Ethernet device is no longer in use.

 

Question 9: You can connect but cannot log on

-User complaint-unable to log on to the server to be connected
-Symptom-the problem usually occurs after a new installation or reset of the workstation, but only one workstation

-Find the problem:

Disconnect the workstation, connect to the network tester, perform IPXPing, obtain the server list, and then check which layer of network protocol is configured. Note the encapsulation type of frames (such as 802.2 ). Now you can test the NIC and check whether the installation and running programs are correct. Check the network adapter's bound frame type (such as 802.3 ). The two frame types must match. If the server and workstation run different frame types, they cannot connect to the Internet.

-Solve the problem:

Use Editor to open the NET. CFG file of the workstation and modify one of the lines: Change FRAME-ETHERNET-802.3 to FRAME-ETHERNET-802.2, connect the workstation to the Internet, and restart. Check the server connection status.

Question 10: Inexplicable intermittent connection interruption

-Support complaints-the service speed is too slow and the connection is intermittently interrupted
-Symptom-the problem only occurs on another website and is irrelevant to the proxy server being accessed.

-Find the problem:

Disconnect the website, connect to a test tool with the function of sending traffic (this is part of the diagnostic process), check the network utilization, collision and error frame level, and determine whether the network is healthy. If the detection is successful but the website is disconnected, it indicates a typical physical layer fault. Perform a cable link test (including user-connected cables). The length test shows that the cable is too long (the UTP cable is 100 meters and the copper axis cable is 185 meters ). Exceeding the allowable length specified by IEEE will cause serious attenuation of transmission signals in the LAN.

-Solve the problem:

Add a repeater to the link to limit the length of the cable to a permitted range. Another method is to use optical fiber instead of cables or copper axes.

Conclusion: from the above examples, we can clearly see that it is necessary to configure an appropriate network testing tool to find common network errors.