7 tips for managing your maintenance network
1. Switch settings
The switch port is configured to 100M full-duplex. The server installs an Intel 100M EISA network card. When the data is transmitted during heavy traffic load, the speed becomes extremely slow. Finally, the network card is found to not support full-duplex. After changing the switch port to half duplex, the fault disappears. This means that the port and the NIC's rate and duplex mode must be the same. At present, there are many adaptive network cards and switches. Due to brand inconsistency, the full-duplex mode cannot be correctly implemented, and only manual forced setting can be solved.
2. Wire sequence of twisted pair
Change the distance between the server and the switch from 5 meters to 60 meters. The result is not connected anyway. Why? Ethernet usually uses two pairs of twisted pairs, arranged at 1, 2, 3, and 6. If you are not using Two pairs of lines, but separate the lines used by the original pair, will form a entanglement, resulting in greater crosstalk (NEXT), affecting network performance. The reason for the above failure is that the pairing lines are not used in 3 and 6, and the connection cannot be made when the distance becomes long. After the RJ45 head was re-wired, everything returned to normal.
3. Network and hard disk
The bottleneck of the network based on file access and printing is the speed of the server hard disk, so configuring the server hard disk plays a decisive role in the performance of the network. Here are a few comments for your reference:
· Select SCSI interface and high speed hard disk.
· The hard disk array card can greatly improve the read/write performance and security of the hard disk. It is recommended.
• Do not allow low-speed SCSI devices (such as CDs) to share the same SCSI channel with the hard disk.
4. Network segment and traffic
A server has two workstations that read and write files very frequently. When the server only installs one network card and forms a separate network segment, all devices on this network segment react slowly. When the server installs two network cards, two are formed. After the network segment, the two workstations with very frequent reading and writing of files are connected to different network segments, and the response speed of all devices in the network has increased significantly. This is because the increased network segment shares the originally concentrated data traffic, which increases the network's response speed.
5. Bridging and routing
Install a microwave networking device. When debugging the Internet, the server always prompts that the current network segment number should be the network segment number of the other party. After the server's network segment number is changed to the other party's, the server's alarm disappears. Ah! It turned out to be a set of devices with bridging properties. Later, I installed a microwave networking device with another location, switched to another manufacturer's product, and then connected, changing the network segment numbers on both sides to be consistent. After installing the device, the server has an alarm: the current routing error. After modifying the network segment on one side, the alarm disappeared. Obviously this is a set of devices with routing properties. The characteristics of the bridge are on the same network segment, and the routes must be on different network segments.
6. Broadcast interference
The above two ends of the networking through the bridge device respectively have a set of application software for transmitting information by broadcasting. When they are running at the same time, both servers will issue an alarm: an incomplete packet is received. After the set of application software is transferred to another network segment, the alarm disappears. This is because the broadcast of the network is unlimited on the same network segment. The two broadcasts interfere with each other to generate an alarm. After moving an application software to another network segment, it is equivalent to routing the broadcast of this network segment and the broadcast on another network segment, thereby limiting the interference of the broadcast, which is also the most important role of the router.
7.WAN and ground
It is not intended to plug the router's power plug into the mains socket. As a result, the 64K DDN cannot be connected. The telecommunications bureau came to check the line is normal, and finally check the grounding voltage of the router power supply, found that it is not right, switch back to the UPS socket, everything is back to normal.
The grounding end of the power plug of the router is broken, which causes the data packets to be lost frequently. When doing PING connection, it is good or bad. Everything works fine after replacing the power cord. Because the WAN connection involves remote lines, the grounding requirements are stricter to ensure strong anti-interference and reach the specified connection rate, otherwise strange faults will occur.
Richerlink specializes in the development and production of network equipment for more than 10 years.
The switch port is configured to 100M full-duplex. The server installs an Intel 100M EISA network card. When the data is transmitted during heavy traffic load, the speed becomes extremely slow. Finally, the network card is found to not support full-duplex. After changing the switch port to half duplex, the fault disappears. This means that the port and the NIC's rate and duplex mode must be the same. At present, there are many adaptive network cards and switches. Due to brand inconsistency, the full-duplex mode cannot be correctly implemented, and only manual forced setting can be solved.
2. Wire sequence of twisted pair
Change the distance between the server and the switch from 5 meters to 60 meters. The result is not connected anyway. Why? Ethernet usually uses two pairs of twisted pairs, arranged at 1, 2, 3, and 6. If you are not using Two pairs of lines, but separate the lines used by the original pair, will form a entanglement, resulting in greater crosstalk (NEXT), affecting network performance. The reason for the above failure is that the pairing lines are not used in 3 and 6, and the connection cannot be made when the distance becomes long. After the RJ45 head was re-wired, everything returned to normal.
3. Network and hard disk
The bottleneck of the network based on file access and printing is the speed of the server hard disk, so configuring the server hard disk plays a decisive role in the performance of the network. Here are a few comments for your reference:
· Select SCSI interface and high speed hard disk.
· The hard disk array card can greatly improve the read/write performance and security of the hard disk. It is recommended.
• Do not allow low-speed SCSI devices (such as CDs) to share the same SCSI channel with the hard disk.
4. Network segment and traffic
A server has two workstations that read and write files very frequently. When the server only installs one network card and forms a separate network segment, all devices on this network segment react slowly. When the server installs two network cards, two are formed. After the network segment, the two workstations with very frequent reading and writing of files are connected to different network segments, and the response speed of all devices in the network has increased significantly. This is because the increased network segment shares the originally concentrated data traffic, which increases the network's response speed.
5. Bridging and routing
Install a microwave networking device. When debugging the Internet, the server always prompts that the current network segment number should be the network segment number of the other party. After the server's network segment number is changed to the other party's, the server's alarm disappears. Ah! It turned out to be a set of devices with bridging properties. Later, I installed a microwave networking device with another location, switched to another manufacturer's product, and then connected, changing the network segment numbers on both sides to be consistent. After installing the device, the server has an alarm: the current routing error. After modifying the network segment on one side, the alarm disappeared. Obviously this is a set of devices with routing properties. The characteristics of the bridge are on the same network segment, and the routes must be on different network segments.
6. Broadcast interference
The above two ends of the networking through the bridge device respectively have a set of application software for transmitting information by broadcasting. When they are running at the same time, both servers will issue an alarm: an incomplete packet is received. After the set of application software is transferred to another network segment, the alarm disappears. This is because the broadcast of the network is unlimited on the same network segment. The two broadcasts interfere with each other to generate an alarm. After moving an application software to another network segment, it is equivalent to routing the broadcast of this network segment and the broadcast on another network segment, thereby limiting the interference of the broadcast, which is also the most important role of the router.
7.WAN and ground
It is not intended to plug the router's power plug into the mains socket. As a result, the 64K DDN cannot be connected. The telecommunications bureau came to check the line is normal, and finally check the grounding voltage of the router power supply, found that it is not right, switch back to the UPS socket, everything is back to normal.
The grounding end of the power plug of the router is broken, which causes the data packets to be lost frequently. When doing PING connection, it is good or bad. Everything works fine after replacing the power cord. Because the WAN connection involves remote lines, the grounding requirements are stricter to ensure strong anti-interference and reach the specified connection rate, otherwise strange faults will occur.
Richerlink specializes in the development and production of network equipment for more than 10 years.
评论
发表评论