Efficient LAN maintenance (fault + + optimization)

Efficient LAN maintenance (fault + + optimization)-Linux Enterprise Application-Linux server application information. For details, refer to the following section. Check the NIC first when there is a fault

In the LAN, network connectivity often occurs. in the case of such problems, we should first carefully check whether Nic settings are normal on the connected machines. During the check, we can open the "control panel/system/device management/network adapter" setting window with the mouse, check whether there are any interruptions and I/O address conflicts in this window (it is best to set the interruptions of each machine to the same for comparison ), until "the device is running properly" appears in the network adapter attributes, and at least you can find yourself in "Network neighbors", it indicates that the NIC configuration is correct.


　　 Confirm that the network cable and network device are working properly

When we check that there is no problem with the NIC, we can look at other computers in the network through the network neighbors. If we still cannot see other machines in the network, this problem may be caused by network disconnection. Network connection failure usually includes network line internal fracture, twisted pair wire, RJ-45 crystal head bad contact, or network connection equipment quality problems, or connection problems. At this time, we can use a line Tester to check whether the line is broken, and then use an alternative method to test whether the quality of network equipment is correct. When neither the network cable nor the network card is faulty, let's see if it is the cause of software configuration. For example, if the interrupt number is incorrect, the fault may occur.


　　 Check if the driver is in good condition

After checking and confirming the hardware, check whether the driver itself is damaged. If not, check whether the installation is correct. If these conditions are normal, the device does not conflict, that is, the network cannot be connected. In this case, you can delete the network adapter in the System Configuration and restart the computer, the system will detect the existence of the new hardware, and then automatically find the driver for installation, during the installation of Windows 98 peer network, I used this method multiple times to solve the problem of "no access to the network.



　　 Set the NIC correctly

If the network media is correct but cannot be connected, return to the NIC settings. Check whether there are device resource conflicts. If there are many conflicts, there are not all prompts. Possible device resource conflicts include: The NE2000 compatible nic and COM2 both use IRQ3 and (Realtek RT8029) PCI Ethernet NIC and display card "like" IRQ10. To solve the conflict, follow these steps: first, block COM2 in the settings window and forcibly set the NIC interruption to 3; if there is a conflict between the NIC and the video card of the PCI interface, we can choose not to Assign IRQ to the display card, that is, set Assign IRQ for VGA in CMOS to "Disable ".
　　

Disable the NIC's PnP Function

Although some NICs support the PnP function, they do not work well or even work well after being installed. To solve this problem, we can shield the NIC's PnP function. To disable the NIC's PnP function, you must run the NIC setup program (generally in the driver package ). After starting the setup program, go to the settings menu. Disable the PnP function of the NIC and change the IRQ that can be set to a fixed value. Save the setting and exit the setting program. If no other device occupies the IRQ, no IRQ conflict occurs. In addition, to install the Windows 95/98 operating system, ensure that the operating system does not treat the corresponding interrupt type as IRQ with the PnP function, therefore, in CMOS, change the interrupt type from PCI/ISA PnP to Legacy ISA ". This method can be used to resolve device conflicts among most PnP NICs, but not necessarily for all PCI NICs, because some Nic setup programs do not provide the option to disable the PnP function.


Reasonably set the server's hard disk

Users who use LAN offices often use the network to print materials and access files. For some reason, the speed of network access may be abnormal. At this time, we often mistakenly think that the cause of reduced network speed may be the bottleneck of some devices in the network, for example, Nic, vswitch, and hub, the speed of the server hard disk is the biggest impact on the network speed. Therefore, correctly configuring the server hard disk in the LAN will greatly improve the network performance of the LAN. We usually need to consider the following factors when setting the hard disk:
The hard disks on the server should be as fast as possible, and the capacity should be large. Because the hard disk speed is fast, the faster the data on the server can be accessed through the network;
The hard disk interface on the server is preferably a SCSI model, because this interface is faster than the IDE or EIDE interface in data transmission. It uses the parallel data transmission mode to send and receive data;
If conditions permit, we can install a hard disk array card for the network server, because the hard disk array card can greatly improve the read/write performance and security of the hard disk;
Of course, we should also note that, in the same SCSI channel, do not share low-speed SCSI devices (such as CD) with hard disks, otherwise, the high-speed data transmission performance of the SCSI interface on the hard disk will not be realized.
　　

Correct use of bridge Devices
　　
A bridge device is a network device used for the same CIDR Block, while a router is a network device used for different CIDR blocks. The author's organization has installed a set of microwave network devices. After the physical devices are connected, the server always prompts that the current network segment number should be the network segment number of the other party. After the network segment number of the server is changed to the same as that of the other party, the server's alarm disappears. Ah! It turns out to be a set of bridging devices. Later, we installed microwave network devices in another location and used products from another vendor. Before the connection, we changed the network segments on both sides to the same one. After the devices were installed, an alarm is reported on the server: The current route is incorrect. After the network segment on one side is modified, the alarm disappears. Therefore, it is important to set network parameters to distinguish between "routing" and "bridge" Devices correctly.
　　


Connect by rule
　　
Each computer connected to the LAN is implemented by twisted pair wires, but they are not simply connected to each other by means of twisted pair wires, we must follow certain connection rules. I tried to connect two computers 100 metres apart with twisted pair wires to achieve communication, but no matter how hard it was, the connection was successful. Later, I was instructed by experts, the connection distance of twisted pair cannot exceed 100 meters. In addition, if we need to connect two computers over 100 meters, we must use a conversion device. When connecting a conversion device and a switch, we must also perform a jumper. This is because in Ethernet, two pairs of twisted pair wires are usually arranged at the positions 1, 2, 3, and 6. If not, instead, the cables used by the original matching pairs are used separately to form a string loop, resulting in a large crosstalk (NEXT ). It has a great impact on network performance. The 10 m network environment is not obvious. In a m network environment, if the traffic is large or the distance is long, the network cannot be connected.
　　

Strictly enforce grounding requirements
　　

Because some weak signals are transmitted in the LAN, interference information may occur in the connection network if the operation is slightly improper or fails to be performed according to the specific operation requirements of the network device, serious problems may cause the entire network to fail. In particular, some network switching equipment, due to the remote connection, have very strict requirements on the ground. Otherwise, the network equipment will not reach the specified connection rate, in this way, various inexplicable failures occur during the network connection process. I have no intention of inserting the power plug of the router into the mains outlet. As a result, the 128 K ddnleased line cannot be connected to the Internet. People from the telecom Bureau checked that the lines were normal. Finally, they checked the zero-ground voltage of the router power supply and found that they were not correct. They switched back to the UPS outlet and everything went back to normal. On the other hand, the power plug of the router is broken to the ground, causing frequent loss of data packets. During PING connection, the connection is always good and bad, and everything works normally after the power cord is changed. It can be seen that when using a network device, it must be performed under the conditions specified by the device. Otherwise, it will bring a lot of trouble to our work.
　　


Use a new network card with good quality and fast speed
　　

In a LAN, it is normal that computers cannot communicate with each other. There may be many causes of failures. According to my previous statistics, most of the failures in the LAN are related to the NIC, or the NIC is not correctly installed, or the network connection is poor, or the NIC is old and cannot be correctly recognized by the computer, in addition, some NICs are installed on servers, which cannot withstand the impact of large data volumes and are eventually scrapped. Therefore, in order to avoid the above phenomenon, we must be willing to invest. If the network card is installed on the server, we must use a good network card, because the server generally runs continuously, only good quality NICs can "work" for a long time. In addition, because the server's data transmission capacity is large, the capacity of the NIC we purchased must match, in this way, we can achieve "good horse and good saddle ".
　　


Vswitch Configuration
　　
A switch is an important data exchange device in a LAN. Correct and reasonable use of the switch can also improve the data transmission performance in the network. I once configured the switch port to m full duplex, and the server was installed with a model of Intel100M EISA Nic. After the installation, everything works normally, but the speed became extremely slow during data transmission at large traffic loads, finally, we found that this Nic does not support full duplex. After the switch port is changed to half-duplex, the fault disappears. This indicates that the speed and duplex mode of the switch port must be consistent with that of the NIC. At present, there are many adaptive NICs and switches. According to the principle, they should be able to correctly adapt to the speed and duplex modes. However, due to brand inconsistency, full duplex modes are often not implemented correctly. It is clear that the server Nic is set to full duplex, but the switch duplex light is not on, only manual forced settings can be solved. Therefore, when setting network device parameters, we must refer to the network device parameters on the server or other workstation to try to make the device match.