Friday, March 18, 2011
Weak current College】 【switch malfunction of the general classification and troubleshooting steps】.
<br> Switch superior performance and a significant drop in prices, led to the increasing popularity of the switch. The network administrator at work often encounter a variety of switch failure, how to quickly and accurately detect the failure and troubleshooting? this article in respect .of common fault type and troubleshooting steps for a brief introduction. Because of the switch in the corporate network application range is extensive, from low-end to the mid-tier, from midrange to high-end, almost every level of the product, so .switch the chance of failure than a router, such as a hardware firewall is much higher, which is why we first discuss switch failure categories and troubleshooting steps. <br> <br> First, switch failure classification: <br> <br> Switch .fault in General can be divided into hardware failures and software faults. Hardware failures are mainly refers to switch the power supply and backplane, module, port and other parts of the fault can be divided into the following categories. <br> <br> ( .1) power failure: <br> <br> Because the external power supply is unstable, or power line aging or lightning, and other causes corrupt or fan stopping power, and thus does not work properly. Because of the power that cause damage to .other parts within things often happen. <br> <br> If the POWER led on the Panel is green, that is normal; if the light is off, the switch does not have the normal power supply. This problem is very easy to find ., is also very easy to solve, but also the most easily preventable. <br> <br> For such a failure, first of all should do the external power supply, usually by introducing separate power line to provide independent power source, and add .the regulator to avoid transient-voltage or low voltage. If conditions allow, you can add a UPS (uninterruptible power supply) to guarantee the normal power switch, UPS provides voltage function, and some not, pay attention to when choosing. In the engine .room set professional lightning protection measures, to avoid damage to the switch and lightning. There are so many do lightning engineering professional company, implementation of network wiring can be considered. <br> <br> (2) port failure: <br> < .br> This is the most common hardware failures, fiber optic ports or twisted pair RJ-45 port, plug connector must be careful. If you accidentally fibre plug dirty, can cause fiber port pollution rather than normal traffic. We often see many people like .to electrification-plug connector, in theory, Yes, but it also inadvertently increase the port's failure rate. When you are not careful handling, can also cause physical damage to the port. If you purchase Crystal head size too large, into the switch, .you can easily destroy the port. In addition, if you add in the port of twisted pair some exposed outdoor, in case the cable were struck by lightning, will result in even the switch port is striking, or cause more unpredictable damage. <br .> <br> In General, the port is a fault or damage several ports. Therefore, when troubleshooting a port connected computer failure, you can replace the connection port, to determine if it is damaged. Experience this type of failure, you can turn .off the power supply, with alcohol tampon cleaning port. If the port is corrupted, it can only replace the port. <br> <br> (3) module fault: <br> <br> The switch is made up of many modules ., such as: stacking module, management module (also called control module), the extension module. These module failure probability is very small, but when problems arise, they will suffer huge economic losses. If you swap modules accidentally, or handling switches is .subjected to collision, or power instability, this can cause this type of failure. <br> <br> Of course, above these three modules are external interface, easier identification, you can also use the led on the module to identify the failure. .For example: stacking module has a flat trapezoid-port switch, or a similar to a USB interface. Management module has a CONSOLE port, and network management to connect your computer to facilitate management. If the extension is fiber connections, there will be a .pair of optical fiber interface. <br> <br> To exclude such failure, first make sure the power switch and module, and then examine the normal supply of each module is inserted in the correct location, last check connection module cable properly. When .the connection management module, but it is provided for connection rate, parity, flow control, and other factors. Connect the extension module, you need to check for matching modes of communication, such as: use full-duplex mode or half-duplex mode .. Of course if you confirm the module failed, the solution is only one, that is, you should immediately contact the vendor for a replacement. <br> <br> (4) backplane failure: <br> <br> Switch for the .module is a patch on the back panel. If the environment is wet, damp circuit circuit board, or components due to factors such as high temperature, lightning strike damage will have Board does not work properly. For example: thermal performance is poor or the .environment temperature is too high, causing the tank temperature instigated components burn out. <br> <br> In the external power supply properly, if you switch the internal modules are not working properly, it may be a backplane is broken, this is the .case even if the electrical maintenance engineer, I am afraid it is last resort, the only solution is to replace the back panel. <br> <br> (5) cable fault: <br> <br> In fact, such a failure .in theory, does not belong to switch their fault, but in actual use, cable fault often lead switch system or port does not work, so here are the kinds of failures into switch hardware failure. For example, connectors and plugs are not tight, .cable production order errors, or when no specification, cable connection should use the crosshair but uses a direct line, cable in the two fibre cross connection, the error line connections are causing network loops, etc. <br> <br> From the above .several hardware failure, the engine room environment is not easy to cause a variety of hardware failure, so we're building room, you must first do a lightning and power, the indoor temperature, indoor humidityDegrees, anti-electromagnetic interference, anti-static .environment building for network equipment normal work to provide a good environment. <br> <br> Second, the switch software failure: <br> <br> Switch software failure refers to system and configuration of fault, it can be divided into the following .categories. <br> <br> (1) system error: <br> <br> Switch system is a combination of hardware and software. There is a switch internal read-only memory refreshes, it holds that this switch is required for the .software system. This type of error and our common Windows, Linux, because at that time the design reasons, there is some vulnerability, in appropriate conditions, can cause a switch, laden, packet loss and error packets, and so on. So switch .system offers such as Web, TFTP, and other ways to download and update the system. Of course when you upgrade your system, or an error occurs. <br> <br> For this type of problem, we need to exercise regularly and browse .equipment manufacturers Web site, if there is a new system introduced or new patches, please timely updates. <br> <br> (2) improperly configured: <br> <br> The Starter switch are not familiar with, or as a result .of a variety of switch configuration, administrators often occurs in a configuration switch configuration error. Such as VLAN classification is not correct cause network is disconnected, the port was mistakenly turned off, switches and network mode configuration mismatch. Such a failure is sometimes difficult to .find, require a certain amount of accumulated experience. If you are not sure that there is a problem with the user's configuration, restore factory default configuration, and then a step-by-step configuration. It is best to configure read the instruction before, .this is the network management to one of the habit of. Each switch has detailed installation manuals, user manuals, drill down into each module has a detailed explanation. Because many switch manual is written in the English language, English is not good for the vendor's .engineers consulting and do specific configuration. <br> <br> (3) of password loss: <br> <br> This may be each administrator has experienced. Once forgotten password, to pass through certain steps to recover or reset the system password .. Some are relatively simple, in the switch pressed a button on it. And others require the adoption of certain procedures to resolve. <br> <br> This type of situation in human forgetfulness or switch failure after data loss occurs. <br> .<br> (4) external factors: <br> <br> Due to viruses or hacker attacks, and other circumstances exist, there may be a host to the connected port to send a large does not comply with the packaging rules for packet, .causing the switch processor excessively busy, forward the packet too late, resulting in a buffer overflow produces packet loss. Another scenario is broadcast storm, it not only consume a large amount of network bandwidth, but also will consume large amounts of CPU processing time. .Network if long time was occupied by a large number of broadcast packets, good point on communications will not be able to function correctly, network speed will slow down or crash. <br> <br> A network card or a port fails, are likely .to trigger a broadcast storm. Because of the switch can only split collision domain, but you cannot split the broadcast domain (without dividing VLAN), so when the number of broadcast packets of up to 30% of the total traffic, network transmission efficiency would .have decreased significantly. <br> <br> The whole software failure should be better than a hardware failure is more difficult to find and solve problems, you may not need to spend too much money, but requires more time. The best work in the .usual habit of logging in. Whenever a problem occurs in time symptom record, failure analysis, failure solution, fault classification summary to accumulate your experiences. For example, sometimes in configure, for various reasons, did not have an impact on the network or no .problems found, but maybe a few days later problems will gradually appear. If you have logging, you can think of is a few configuration errors. Because very often ignore this, and thought it was a problem in other respects, when away many detours, .it found the problem. So logging and maintenance information is very necessary. <br> <br> Third, switch failure of General troubleshooting steps: <br> <br> Switch fault varied and different fault has different manifestations. Fault analysis to flexible application .in a variety of methods (such as exclusion from a comparative law, replace, or method) that identify the fault lies, and promptly excluded. <br> <br> (1) exclusion: <br> <br> When we are faced .with the symptom and diagnose problems, inadvertently had already learned to use a process of elimination to determine the orientation of failure. This approach is based on the observed symptoms, as comprehensively as possible list all possible failures, and then analyze and troubleshoot. The exclusion .is to be followed to the propagation of Jane, improving efficiency. By using this method you can meet all kinds of trouble, but attendants need strong logical thinking, comprehensive knowledge on the switch. <br> <br> (2) contrast: < .br> <br> The so-called contrast is to leverage existing, the same model and to the normal operation of the switch as a reference object, and the contrast between the switch failure to identify the point of failure. This method is simple and .effective, especially in the system configuration of fault, simply compare to find the configuration of different points, but sometimes you want to find one of the same model, configure the same switch is not an easy task. <br> <br> (3 .) alternate method: <br> <br> This is the most commonly used method, also in maintenance computer intensive methods. Alternate method is to use a normal switch parts to replace the defective parts may be to identify the point of failure. It is .mainly used for hardware failure diagnosis, but note that replacement parts must be of the same brand, same model of the switch. <br> <br> Of course, in order to enable troubleshooting work rule-based, we can faultAnalysis, in accordance .with the following principles to analyze. <br> <br> 1 from the rear to front <br> <br> Because the switch general applause (eg port failure) are connected to the computer and found, often from the client to initiate check .. We can walk along client computer ->> port module - - horizontal cable ->> jumper switch so that a route, one by one, checking, to exclude the possibility of remote fault. <br> <br> .2, from the outside and inside <br> <br> If the switch is down, we can start on the external lights of all kinds of discrimination, and fault indication, and then to check whether the internal parts of the problem. For POWERLED .green light indicates power supply for normal, off means no power supply; LINKLEDs yellow indicates that now the connection work in 10Mb / s, green indicates to 100 MB / s, off that is not connected, the blinking indicates that the port manually by the .administrator; RDPLED represents a redundant power supply; MGMTLED represents the Administrator module. Regardless of whether the failure from the outside, you must log on the switch to determine specific failures and make the appropriate troubleshooting steps. <br> <br> 3, from .soft to hard <br> <br> In the event of a failure, nobody wanted to go on the screwdriver to remove the switch, so check, always start with the system configuration or system software to proceed with the investigation. If the software does .not resolve the problem, that is, there is a problem with the hardware. For example, a port is not good, then we can check that the user is connected to the port is not in the appropriate VLAN, or that the port is closed .by another administrator, or configuration of other reasons. If you exclude the system and configuration possibilities that can be suspected that the real problem is - the hardware failure. <br> <br> 4, parts difficult to <br> <br> In .the face of failure analysis is more complex, you must first configure from simple operation to proceed to exclude or. This can accelerate the speed of troubleshooting, improving efficiency. <br> <br> Fourth, to summarize: <br> <br> .The wide variety of symptoms due to a switch, there is no fixed rule out steps, while failure often have a clear direction, can recognize at a glance. So only the specific analysis of a case-by-case basis, of course, no .matter what kind of failure for a new network administrators are difficult, so if you want to become a master switch malfunction of, must be in their day-to-day work experience, one issue per ready are hard to review the source of the problem .and solution. In order to constantly improve ourselves better accomplish the important task of the network management. <br>.