The content of the invention
In consideration of it, being necessary to provide a kind of server failure detection means and method that can feed back failure cause self-healing failure.
A kind of server failure detection means, including:
One baseboard management controller, the baseboard management controller is used for the execution state of the basic input output system of the reception server, the baseboard management controller is also preset with some presupposed solutions of some fault types of server basic input output system and correspondence some fault types, when the server basic input output system exports Fisrt fault signal, the baseboard management controller identification Fisrt fault signal determines failure for Fisrt fault type and performs corresponding first presupposed solution.
A kind of server failure detection method, comprises the following steps:
Server is started shooting;
The signal of the execution state of the basic input output system of server basic input output system output representative server is to baseboard management controller;
The baseboard management controller judges whether the basic input output system of the server performs state abnormal;
If the basic input output system of the server performs abnormal state, the baseboard management controller is called and performs solution corresponding with abnormality.
The server failure detection means can meet with assisting user when failure cannot light screen in BIOS and understand failure cause and automatically attempt to repair failure.
Specific embodiment
Fig. 1 is refer to, server failure detection means of the present invention 10 is applied in a server 100.The server failure detection means 10 includes baseboard management controller 11.The baseboard management controller 11 is connected with the basic input output system 101 of the server 100.The baseboard management controller 11 is used for the execution state of the basic input output system 101 of the reception server 100.The baseboard management controller 11 is also preset with some presupposed solutions of some fault types of the basic input output system 101 of server 100 and correspondence some fault types.When the basic input output system 101 exports Fisrt fault signal, the identification Fisrt fault signal of the baseboard management controller 11 determines failure for Fisrt fault type and performs corresponding first presupposed solution.
In present embodiment, the baseboard management controller 11 is preset with the fault type of the several basic input output system 101.The fault type includes but is not limited to cabinet and invades failure, CPU initialization failures, cpu frequency setting failure, cpu cache initialization failure, VBIOS initialization failures, internal memory initialization failure, memory size failure, hard disk initialization failure, PCI external equipment failures, USB external equipment failures, VBIOS collapse failures, platform controller initialization failure, node administration controller failure etc..
In present embodiment, when the identification failure of the baseboard management controller 11 is that cabinet invades failure, the baseboard management controller 11 calls the first presupposed solution, the baseboard management controller 11 confirms whether the cabinet is installed correctly, if cabinet installs correct, the baseboard management controller 11 is removed the cabinet and invades failure logging to continue starting procedure.
When the identification failure of the baseboard management controller 11 is CPU initialization failures, depositor related setting value is inserted depositor according to 100 depositor setting table of server by the baseboard management controller 11, and is outputed signal to CPU and restarted pin control CPU and restart.
When the identification failure of the baseboard management controller 11 is that cpu frequency sets failure, the baseboard management controller 11 inquires about now cpu frequency setting, if setting is abnormal reads cpu frequency from the mainboard CMOS ROM of server 100.
When the identification failure of the baseboard management controller 11 is cpu cache initialization failure, the baseboard management controller 11 inquires about the cpu cache setting of the server 100, if setting is abnormal resets caching.
When the identification failure of the baseboard management controller 11 is VBIOS initialization failures, the baseboard management controller 11 determines whether display card, if there is no display card, then video frequency output set of options is by CPU output images by the baseboard management controller 11, if there is display card, display card is restarted.
When the identification failure of the baseboard management controller 11 is internal memory initialization failure, the ROM read on internal memory obtains memory standards information, and the specification information is compared with 100 setting value of server, the setting value of server 100 is revised as if having differences the memory standards information for obtaining.
When the identification failure of the baseboard management controller 11 is memory size failure, the ROM that the baseboard management controller 11 is read on internal memory obtains memory standards information, and the specification information is compared with 100 setting value of server, the setting value of server 100 is revised as if having differences the memory standards information for obtaining.
When the identification failure of the baseboard management controller 11 is hard disk initialization failure, the baseboard management controller 11 checks the controller setting of hard disk.
When the identification failure of the baseboard management controller 11 is PCI external equipment failures, device PCI is scanned and by abnormal device PCI information output to the baseboard management controller 11.
When the identification failure of the baseboard management controller 11 is USB external equipment failures, scans USB device and abnormal USB device information is exported to the baseboard management controller 11, this USB device is disabled and reinitialized.
When the identification failure of the baseboard management controller 11 is that VBIOS collapses failure, the basic input output system 101 reads VBIOS from standby ROM.
When the identification failure of the baseboard management controller 11 is platform controller failure, basic input output system 101 detects whether the platform controller has feedback to judge whether hardware is normal.
When the identification failure of the baseboard management controller 11 is Node Controller failure, the baseboard management controller 11 is tested to Node Controller using IPMB interfaces, and if abnormal, 11 pairs of Node Controllers of the baseboard management controller carry out software upgrading.
Fig. 2 is refer to, the better embodiment of server failure detection method of the present invention 200 includes step S10-S16:
S10:Server is started shooting;
S12:The signal of the execution state of the basic input output system of server basic input output system output representative server is to baseboard management controller;
S14:The baseboard management controller judges whether the basic input output system of the server performs state abnormal;
S16:If the basic input output system of the server performs abnormal state, the baseboard management controller is called and performs solution corresponding with abnormality.
Finally it should be noted that, above example is only to illustrate technical scheme and unrestricted, although being described in detail to the present invention with reference to preferred embodiment, it will be understood by those within the art that, technical scheme can be modified or equivalent, without deviating from the spirit and scope of technical solution of the present invention.