No Network Communication Right After the Switch Is Powered On
[Symptom]
When the switch has just been powered on, it cannot connect to other networks and requires a waiting period before it can. In addition, only after it has been in use for some time does access to other computers become fast; if the network is left unused for a while, access slows down again.
[Analysis]
Because this is a managed switch, and to prevent a topology loop in the network from bringing it down, managed switches enable Spanning Tree Protocol by default. This ensures that even if a loop exists, only one path is retained while other links are automatically cut off. Therefore, when the switch powers up, each port must pass through the listening, learning and forwarding states in turn, a process that takes about 3 to 5 minutes. If the switch must start up quickly, you can enable “PortFast” on the ports connected directly to computers, so that those ports immediately and permanently switch to the forwarding state. This lets the devices connect to the network right away and avoids the wait required for ports to transition from the listening and learning states to the forwarding state.
[Solution]
If data forwarding must begin immediately after the switch is powered on, you can disable Spanning Tree Protocol or set the ports to PortFast mode. Note, however, that although these two methods skip the port detection process, once a topology loop forms between network devices, it will bring network communication down. A 5-port switch can only use 4 ports
[Symptom]
An office has 4 computers but only one information outlet, so a 5-port switch (one port of which is an UpLink port) was deployed. The assumption was that the 4 computers would occupy 4 ports and the single UpLink port would connect to the LAN, but after connecting to the network, port 1, which is adjacent to the UpLink port, could not be used normally.
[Analysis]
The UpLink port cannot be regarded as a separate port, because it is in fact the same port as the adjacent one; only the connection target it serves differs. With the UpLink port, a hub device can use a straight-through cable to connect to the ordinary port of another hub device, so a crossover cable is unnecessary. The chips in switches and hubs are usually x4, so hub devices generally have 4, 8, 16 or 24 ports; building one with 5 ports would waste 3 modules and thus increase cost.
[Solution]
Replace the 4-port switch with an 8-port switch and the problem is resolved.
The “COL” indicator stays on or keeps flashing, and communication cannot be established
[Symptom]
Computers on the LAN access the server through a hub, but one day all client computers were unable to connect to the server, and Ping between client machines was intermittent. Inspection of the hub revealed that the “COL” indicator was either steadily on or continuously flashing.
[Analysis]
The “COL” indicator shows collisions and conflicts in the network. A flashing “COL” light indicates that collisions are occurring; a steadily lit “COL” light indicates that a large number of collisions are occurring. The cause of such heavy collisions may be a hub fault or a NIC fault. Generally, a NIC fault is less likely, so the focus was placed on troubleshooting the hub.
[Solution]
After replacing the hub, the network returned to normal. After upgrading to a gigabit network, the server connection is intermittent
[Symptom]
The server originally used a 10/100Mbit/s NIC and ran perfectly well. But after a 1000Mbit/s NIC was installed and connected to the 1000Base-T port of the central switch, the server’s connection to the network became intermittent and extremely unstable, and it could not provide normal network services. Testing the network with a cable tester showed no connectivity problem on the twisted-pair link.
[Analysis]
The connection worked normally at 100Mbit/s and the fault appeared only after upgrading to 1000Mbit/s, so the cause would appear to be a Category 5e cabling issue. Although in theory a Category 5e system supports a transmission rate of 1000Mbit/s, if the twisted-pair cable, patch panel, network cable or other network equipment is not of good quality, or if the termination workmanship is problematic, 1000Mbit/s bandwidth still cannot be achieved. Because 1000Base-T requires all 4 pairs of the twisted-pair cable, with an effective transmission rate of 250Mbit/s per pair and full-duplex transmission, it places higher demands on the electrical performance of the twisted pair, including signal attenuation, echo, return loss, crosstalk and resistance to electromagnetic interference. If the twisted-pair cable or other components perform poorly, severe crosstalk will occur between wire pairs, causing communication to fail.
[Solution]
Given that the performance of the Category 5 cabling system might not meet the needs of a gigabit network system, the problem was resolved by switching to Category 6 cabling products. Although the Link light keeps flashing, the network speed is extremely slow
[Symptom]
The server’s Internet speed was very slow. At first, web pages opened very slowly; later, pages could not even be opened, and Pinging websites could not resolve addresses. It was initially assumed to be a DNS setting or a server fault, but these were all running normally. Pinging other computers revealed a very high packet loss rate. At the same time, the switch’s Link indicator kept flashing and data was being exchanged very frequently, indicating that the computer was continuously sending and receiving packets. Restarting the switch relieved the problem, but it reappeared after a while.
[Analysis]
Judging from the symptoms, this is a broadcast storm within the network. Broadcast storms can have many causes, such as worm viruses, switch port faults, NIC faults, link redundancy without Spanning Tree Protocol enabled, incorrect network cable wire order, or interference. Checking the switch indicators when a network fault occurs is a very convenient diagnostic method, as it provides an intuitive view of network connectivity and network traffic.
[Solution]
In the current situation, a worm virus is the main cause of the network outage. The problem was resolved by promptly updating the server’s system patches, installing network-based antivirus software, promptly upgrading the server’s virus database, and installing an antivirus client program on the server.
