Hardware breakdown is to show server hardware appears unusual and those who bring about is of all kinds and wrong. Because the server is formed more complex, when because this is in,be being checked must serious, careful. A specification is with a LH6000 below.

Have a HP LH6000, deserve to have 256M memory, use a PIII XEON 700 to take the processor of cache of 2M high speed. There is any indication after switching on the mobile phone, but the information that hinted voltage of a CPU is 0 bend over on systematic daily record, 3 lamps are in systematic indicator light ceaselessly twinkle (indicator light the type of police of another kind of newspaper that 3 lamps twinkle is a server, I can explain after article) . This kind of mistake is processor commonly voltage adjustment module (VRM) make mistake or CPU makes mistake or CPU and CPU board piece the contact is undesirable, but also may be CPU board piece make mistake, at this moment the case is more complex, must pass serious and discreet reflection. Because of CPU board piece in whole server, have decisive place, if it makes mistake,the server can sign up for fatal error, and be in systematic daily record can hint fatal error, but the circumstance that signs up for CPU voltage fault also has 5% to control. We insert CPU exchange in another CPU immediately in groove, the sort of a moment ago breakdown still is after switching on the mobile phone. It is so in preliminary judgement, can eliminating is CPU board piece bad.

At this moment, take out CPU to wipe golden finger carefully, and CPU board piece in after the place that contacts with CPU, switch on the mobile phone to still be not had show.

For the situation with opposite bad processor processor voltage module (VRM) the case of occurrence breakdown is bigger. Take module of next processor voltage in another LH 6000 instantly then, installation is in this server. After switching on the mobile phone, the server still does not have any indication, still hint CPU voltage is the information of 0 bend over on systematic daily record, 3 lamps still are in systematic indicator light ceaselessly twinkle. At this moment the case is clearer. After taking next CPU to install from inside another LH6000 instantly then, it is normal to switch on the mobile phone.

In the maintenance of the server, clew can appear complicated and confusing, judge the seat that gives an issue well and truly impossibly generally speaking. Ask relevant personnel wants to have confidence and patience so. The flow with occurrence common error is to pass the information on systematic daily record to solve, if did not solve a problem to find out other element again, watch log news again next. Anyhow, after the server makes mistake, must one pace is solved, without shortcut but character.

Be like again:

A HP LH 4 switchs on the mobile phone not to show, daily record of the system when discovery switchs on the mobile phone does not have any information, and systematic indicator light is not bright. Preliminary judgement is power source respect appeared mistake. Through microscope, the power source that discovers a server is normal, because this the largest possibility is the power source of the server,administrative board appears breakdown. Change power source management board hind, switch on the mobile phone it is normal to show. But at this moment, new issue came: When self check, use CTRL M to cannot detect to hard disk.

Hard disk is normal on other server, because this keeps clear of instantly the CMOS of this server, but still abnormal. I establish engrave net to find the newest BIOS of this server, upgrade BIOS hind is insoluble also problem. After checking hard disk basket and the data line in the server and power supply cord again, still make mistake. At this moment, general situation can suspect the I/O that is a server board (input output board piece) have a problem. But be in this moment, I send the old-style net that there is HP of a blame on I/O board now to get stuck, instantly the server after this net blocks purify with respect to everything normal.

Hardware breakdown does not show hardware has a problem only, it also points to hardware between incompatible. Because of the server run need to be coordinated between each component energetically normally. When suggesting everybody is purchasing each component, use same brand to be installed formerly, the component that and should be used can develop server property (go up although a medium old-style net card also is met normally function of serious effect server) , such ability won't produce the breakdown of be baffled.

I ever had encountered a kind of situation: The user needs to upgrade his HP LH6000 to double net card, I suggest he buys the former check that install a network, but after the INTEL 82559 chip that sees when him the net card of HP LH6000 is to use, the net that decides not to use former outfit to the net gets stuck and use another brand to also use INTEL 82559 absolutely gets stuck. Passed a few days, he phones me to say, his new network card cannot use a network to reach data desired result redundantly, suspect the server has a problem. I took card of net of a HP to arrive over there the user, microscope after the environment of the server is completely normal, everything is normal after installing HP net check to the machine. This example more explained, want to develop the greatest performance of the server and function, must use a replacement that former brand installs formerly. Be not former brand to be not the fittings of former outfit, cannot support the certain function of the server, use what can affect a server badly normally.

Generally speaking the server of medium, high end calls the police the system is more perfect, besides systematic daily record, still have indicator light. With HP LH6000 for, the green lamp Chang Liang of indicator light states the server is normal; Green light shines and yellow twinkles state the server has trouble, but not be deadly; If 3 lamps twinkle (green, yellow, red 3 lamps) state the server has critical failure, the server stops to move. Be compared and character, indicator light can hint only more general breakdown, and systematic daily record is more complete. In maintenance, must look carefully at these two kinds carefully to call the police systematic information. Having what must notice is systematic daily record it is a memory, the capacity is finite (LH 6000 can store 200 information) . When the capacity is insufficient must clear sky, otherwise the server will call the police, it is server indicator light commonly the newspaper is not fatal error, but cannot store any information again however.

Comments

No Responses to “Server hardware breakdown handles a few”

Leave a Reply