CN106598790A - Server hardware failure detection method, apparatus of server, and server - Google Patents

Abstract

The present invention provides a server hardware failure detection method and its device and server, the method includes detecting that the server's BIOS device enters the start-up phase; Fault detection and analysis is performed on the hardware in each working stage, and the working stage includes the startup stage; the hardware fault information detected by the BIOS device covers the entire cycle of the server's operation, and the fault pre-detection is carried out on the hardware of the server, thereby Timely processing the failure of the server during operation improves the stability and reliability of the server operation; further, the basic input and output system device stores the hardware failure information obtained by detection and analysis, which is convenient for personnel to deal with the failure, And the unified storage management of the hardware fault information is realized.

Description

A server hardware fault detection method and its device and server

technical field

The invention relates to the field of computers and communications, in particular to a server hardware failure detection method and its device and server.

Background technique

In the current mid-to-high-end servers, the server generally has some "black box" functions, which are used to record fault information when the operating system crashes, and can record various kernel exceptions of the OS (operating system, Operating System) such as kernel errors, restarts, and resets. , Abnormal print information, etc. can also be recorded through SEL (System Event Log, System Event Log) to record some simple hardware errors, or through out-of-band methods (such as joint test links) to collect errors on site after a fault occurs , or passively monitor device abnormalities through the in-band abnormality trigger mechanism, which needs abnormal conditions to trigger its abnormality recording module to record. These methods can help maintenance personnel to determine the cause of the fault to a certain extent, but these methods still have the following defects:

1. The above-mentioned method uses passive trigger detection records, lacks active detection of servers, especially active screening and monitoring of server hardware failures. When the system starts and runs normally and the service quality drops significantly, the system does not trigger the fault information record, which will cause the fault information to be missed, making it difficult for maintenance personnel to trace the fault information during maintenance.

2. Since the fault information will be recorded for detection only when the system crashes or an abnormal trigger occurs, the ability to collect and analyze hardware faults during system (business) operation is seriously insufficient, resulting in the early warning capabilities of the system Insufficient, reducing the stability and reliability of the system.

3. The recorded fault information is too simple and scattered. Without accurate and unified record management, it is impossible to analyze the fault information in one step. In the later stage, a lot of analysis, screening and cross-validation are required to find the main fault source.

4. The collection of fault information through out-of-band will be limited by professionals, site environment, information security, etc., and the cost of environment deployment, personnel coordination, and environment recovery will be high.

Therefore, in the current implementation scheme of server fault information recording, the detection and recording of fault information can only be realized under specific conditions, and the recorded fault information is simple and scattered, requiring a large amount of analysis in the later stage.

Contents of the invention

The main technical problem to be solved by the present invention is to provide a server hardware fault detection method and its device and server, to solve the technical problem that in the prior art, it is impossible to detect and record and store the real-time fault information of the hardware in each working stage of the server .

In order to solve the above technical problems, the present invention provides a server hardware fault detection method, comprising:

the server's BIOS device detects that the server has entered a start-up phase;

The BIOS device starts to detect the hardware failure of the server in each working stage, and the working stage includes the start-up stage;

The BIOS device stores the detected hardware failure information.

In an embodiment of the present invention, the start-up phase includes an initialization phase, and the BIOS device performs hardware failure detection on the server during the initialization phase, including:

The basic input and output system device performs hardware pre-detection on at least one of the CPU, memory, chipset and power supply of the server according to the hardware detection mechanism provided by the server to obtain current hardware information, and obtains current hardware information from the hardware information. The faulty hardware information is screened out for analysis and processing to obtain corresponding hardware fault information.

In another embodiment of the present invention, the startup phase further includes a device enumeration phase, and the BIOS device performing hardware failure detection on the server during the device enumeration phase includes:

The BIOS device obtains the status information and resource information of each hardware on the server, and identifies the fault information of the faulty hardware therefrom.

In another embodiment of the present invention, the startup phase is a cold startup phase or a hot startup phase.

In another embodiment of the present invention, the working phase further includes at least one of an operating system pre-booting phase and an operating system service running phase.

In another embodiment of the present invention, when the working stage includes an operating system pre-boot stage, the BIOS device performing hardware fault detection on the server during the operating system pre-boot stage includes:

The BIOS device pre-tests the out-of-band hardware device of the server to be booted;

Obtain current hardware information of the hardware device;

Screen out the fault information of the faulty hardware device from the current hardware information;

When the working stage includes the operating system service running stage, the BIOS device performing hardware fault detection on the server during the operating system service running stage includes: the basic input output system device judging the hardware failure of the server Whether an interrupt signal arrives, and if so, the BIOS device detects the relevant hardware of the operating system; obtains the fault information of the hardware.

In another embodiment of the present invention, before the BIOS device stores the detected fault information, it further includes allocating a fault on the server serial flash memory for storing the hardware fault information. storage area.

In order to solve the above technical problems, the present invention also provides a basic input and output system device, comprising:

The fault information detection trigger module is used to detect whether the server enters the startup phase;

The fault information detection module is configured to, when the fault detection triggering module detects that the server enters the start-up phase, start to detect the hardware fault of the server in each working phase, and the working phase includes the start-up phase;

The fault information storage module is used for storing the hardware fault information detected by the fault information detection module.

In another embodiment of the present invention, it also includes a storage setting module, configured to allocate a memory on the server serial flash memory for storing the hardware failure information before the failure information storage module stores the hardware failure information. Fault storage area for hardware fault information.

In order to solve the above technical problems, the present invention further provides a server including the above-mentioned BIOS device.

The beneficial effects of the present invention are:

A kind of server hardware fault detection method and its device and server provided by the present invention, when the basic input output system (Basic Input Output System, BIOS) device of the server detects that the server enters the start-up phase, it starts to monitor each working phase of the server The hardware performs fault detection and analysis to obtain corresponding hardware fault information. Since the server's own BIOS device is used, it can detect all hardware failures that may occur during the entire cycle of server operation, which can improve the comprehensiveness and accuracy of hardware failure information detection, and is more conducive to realizing server hardware failure. The unified storage and management of information ensures that maintenance personnel can accurately obtain hardware fault information when maintaining the server, and know the location of the hardware that needs fault handling and the cause of the fault, further improving the stability and reliability of the server .

Description of drawings

Fig. 1 is the flowchart of the server hardware fault detection method provided by the present invention;

Fig. 2 is the flow chart of hardware failure detection in server initialization stage provided by the present invention;

Fig. 3 is the flowchart of hardware failure detection in the equipment enumeration stage of the present invention;

Fig. 4 is the flow chart that the operating system pre-boot phase of the present invention carries out hardware failure detection;

Fig. 5 is the flow chart of hardware fault detection in the operation stage of the operating system of the present invention;

Fig. 6 is a structural block diagram of the basic input and output system device provided by the present invention.

detailed description

The present invention will be further described in detail below through specific embodiments in conjunction with the accompanying drawings.

Embodiment one:

Please refer to FIG. 1. FIG. 1 is a flow chart of the server hardware fault detection method provided by the present invention. It should be understood that the server hardware fault detection method provided in this embodiment is to actively monitor the server hardware through the basic input output system device. Fault detection, active here means that according to the detection mechanism preset by the server, when the server is started, the BIOS device immediately performs a fault detection operation on the running hardware of the server or the various tasks of the BIOS device on the server The hardware fault detection operation is carried out in each stage, which specifically includes the following steps:

S101. The BIOS device of the server detects that the server enters the start-up phase;

In this embodiment, the startup phase of the server is a cold startup phase or a hot startup phase; the detection by the BIOS device that the server enters the startup phase means: when the startup phase is a cold startup phase, the The basic input and output system device can detect whether it enters the startup phase through the following methods but not limited to the following methods: detecting whether the power switch key on the server is pressed or detecting whether the power supply circuit of the server is connected to the server power interface or by checking the status of the power supply flag, if so, the server has entered the start-up phase, the server has been running, and step S102 is performed, otherwise, continue to detect;

When the startup phase is a hot startup phase, the basic input and output system device detects whether the server has a reset startup signal input, if so, the server starts a hot startup operation, and performs step S102, otherwise, continues detection; here The reset start signal can be triggered by hardware, such as: input through the reset button; it can also be input through software, such as: through codes and tools to achieve regular input to the server; it can also be the user actively through the command Or operate the "Restart" button to enter.

S102. The BIOS device starts to perform fault detection on the hardware of the server in each working stage, and the working stage includes the start-up stage;

S103. The BIOS device stores the detected hardware failure information.

In this embodiment, before step S103, it also includes allocating a fault storage area for storing the hardware fault information on the server serial flash memory; further, the BIOS device records and stores The fault information includes: time, event, severity, specific location or fault details, and suggested handling methods.

In this embodiment, after the above-mentioned steps are performed to detect and store the hardware failure information, when the maintenance personnel need to maintain the server, the maintenance personnel can use the out-of-band control platform connected to the storage area or the network The user interface directly obtains the stored hardware fault information, which is convenient for maintenance personnel to track the fault occurrence track, restore on-site, and replace the faulty hardware (for example: directly replace a certain CPU, directly replace the first memory stick, and directly replace the faulty bus interface card) . In middle and high-end servers, the hot-swapping technology (including but not limited to: CPU hot-swapping, memory hot-swapping, and bus interface hot-swapping) can fully guarantee the uninterrupted operation of the system, achieving early detection, early warning, early prevention, purpose of early treatment. Even if the server hangs up during cold start or hot start, and the peripherals of the server cannot be used (for example, the network port is not connected, the screen is not lit, and the keyboard and mouse do not respond), effective fault information can still be obtained.

In this embodiment, the operation and maintenance personnel obtain the hardware fault information through the control platform, and besides processing it in time, they can also dump the hardware fault information to another out-of-band storage device.

In this embodiment, the startup phase includes an initialization phase, and the step of the BIOS device performing fault detection on the hardware of the server during the initialization phase is shown in FIG. 2 , which specifically includes:

S201, the BIOS device initializes CPU, memory, chipset and power supply;

S202. The BIOS device detects and acquires current hardware information of at least one of CPU, memory, chipset and power supply;

In this embodiment, the basic input and output system device performs pre-detection on at least one of the server's CPU, memory, chipset and power supply according to the hardware detection mechanism provided by the server to obtain current hardware information, from Faulty hardware information is screened out from the hardware information for analysis and processing to obtain corresponding hardware fault information.

Specifically, in this embodiment, when the BIOS device detects that the server has entered the initialization phase, the BIOS device can use the BIOS device itself or actively increase the pressure, or use the CPU and chipset The provided hardware detection mechanism, or the use of in-band integrated tools (such as memory test tools, system event log test tools), etc., actively initiates detection of server hardware faults and configurations such as CPU, memory, chipset, and power supply, and obtains Corresponding hardware information, and then perform pre-analysis and judgment, pre-statistics, pre-screening, scanning, and measurement of the hardware information obtained, and collect test results, and filter out effective fault information (including those that may trigger subsequent system exceptions) Information) is recorded and stored in detail; so that when a system abnormality occurs on the server at this stage, it is ensured that the server obtains and records more detailed hardware failure information before the system abnormality occurs. In this stage, the fault information stored in the record includes but is not limited to: CPU errors and warnings, CBO (caching agent, Caching Agent) errors and warnings, QPI (Quick Path Interconnect, QuickPathInterconnect) errors and warnings, IIO (integrated input /Output, Integrated I/O) port errors and alarms, HA (Home Agent, Home Agent) errors and alarms, IMC (Integrated Memory Controller, Integrated Memory Controller) errors and alarms, PCU (Power Control Unit, Power Control Unit) errors Errors and alarms, power supply and voltage errors and alarms, memory errors and alarms (including errors and alarms of the memory module itself, memory channel errors and alarms, memory insertion errors and alarms, memory voltage errors and alarms, memory incompatibility errors and alarms, configuration errors and warnings, etc.).

In this embodiment, the startup phase also includes a device enumeration phase, and the flowchart of hardware fault detection in this phase is shown in Figure 3, specifically including the following steps:

S301. The BIOS device starts device enumeration;

S302. The device of the basic input and output system detects and acquires current information of the device;

Further, the BIOS device obtains the status information and resource information of each hardware on the server, and identifies the fault information of the faulty software and hardware therefrom. At this stage, the fault information includes but is not limited to: device access error (including illegal memory and IO requirements), third-party firmware (OPTION ROM) is not executed (including insufficient space, wrong format), device damage is disabled. Specifically, in this embodiment, when the server sends a probe task to the bus interface (Peripheral Component Interface Express, PCIE) peripherals and calculates resource requirements, the BIOS device starts to identify the first set of industrial specifications according to the detection mechanism. Three-party firmware (OPTION ROM) identifier, manufacturer information, device classification information and capacity, check hardware status indication information (such as link status, bandwidth information, etc.), and identify faulty hardware fault information from the above information for storage .

In this embodiment, the working phase also includes at least one of the operating system pre-boot phase and the operating system business running phase; please refer to Figures 4 and 5, which respectively perform hardware Flow chart of fault detection;

As shown in Figure 4, the hardware failure detection and analysis in the pre-boot stage of the operating system includes the following steps:

S401. The BIOS device performs pre-detection on the out-of-band hardware device of the server to be booted;

S402. Obtain current hardware information of the hardware device;

S403. Filter out the fault information of the faulty hardware device from the current hardware information;

In this embodiment, the out-of-band hardware devices of the server include but are not limited to: hard disks, server network ports, and device boot properties; the fault information includes but not limited to: no bootable devices, hard disk (or U disk) damage (including MBR partition damage), PXE network boot failure (including port information, network ping failure), ME (Management Engine) working status is abnormal. Preferably, at this stage, when the BIOS device performs fault detection on the hard disk partition, the BIOS device actively initiates a detection and acquisition signal to obtain the master boot record ( MBR partition) data, analyze guide sign, end sign and error information data area, judge whether hard disk (U disk) can guide, damage according to the hardware detection machine that described server provides; communication link status and working mode; through DHCP (Dynamic Host Configuration Protocol, Dynamic Host Configuration Protocol) communication to check whether the network is connected; list the single-board boot devices, and check whether there are bootable devices.

As shown in Fig. 5, the hardware failure detection and analysis in the operation stage of the operating system includes the following steps:

S501, judging whether a hardware interrupt signal of the server arrives;

S502, if yes, the BIOS device detects related hardware of the operating system;

S503, acquiring fault information of the hardware;

In the above fault detection and analysis, when it is judged that the hardware interrupt signal arrives, the BIOS device detects the fault of the hardware related to the business operation, and analyzes and classifies the detected hardware fault information , statistics, and then store the fault information. The fault information detected at this stage includes but is not limited to: CPU errors and warnings, CBO errors and warnings, QPI errors and warnings, VT-D errors and warnings, IIO port errors and warnings, memory errors and warnings, PCIE errors and warnings, PCU errors and warnings, Ubox (Utility Box) errors and warnings. Preferably, in the hardware failure detection process at this stage, the BIOS device starts the MCA (Machine Check Architecture) function and the enhanced error record AER (Advance Error Report) function, and opens the error detection block (Machine Check Error Bank) corresponding to each component. ) switch, hook up the fault identification classification function and the error handling hook function of each component. When an MCE (Machine-Check Exception) exception occurs, the hardware pulls down the error status pin to generate a system management interrupt (SMI). Now described BIOS device obtains control right, reads the error state register that CPU and bridge chip carry by hardware failure identification classification function, obtains error detection block (Machine Check Error Bank) concrete information, then carries out detailed analysis according to chip manual, Separate and interpret specific hardware error messages.

Embodiment two:

This embodiment provides a basic input and output system device. It should be understood that the BIOS device can be installed in any server to realize hardware failure detection of the server at any working stage. Please refer to FIG. 6, the basic input and output system Device 60 includes:

A failure information detection trigger module 61, configured to detect that the server enters the start-up phase;

The fault information detection module 62 is configured to start fault detection on the hardware of the server in each working stage, and the working stage includes the startup stage;

The fault information storage module 63 is used for the BIOS device to store the hardware fault information obtained through detection and analysis.

In this embodiment, at the start-up stage of the server, the fault information detection module 62 pre-detects at least one of the CPU, memory, chipset and power supply of the server according to the hardware detection mechanism provided by the server to obtain the current hardware information, and screen faulty hardware information from the hardware information for analysis and processing to obtain corresponding hardware fault information.

In the device enumeration stage of the server, the fault information detection module 62 obtains the status information and resource information of each hardware on the server, and identifies the fault information of the faulty hardware therefrom.

During the pre-boot stage of the operating system of the server, the fault information detection module 62 pre-detects the out-of-band hardware device of the server to be booted;

Obtain current hardware information of the hardware device;

Screen out the fault information of the faulty hardware device from the current hardware information;

During the operation stage of the operating system business of the server, the fault information detection module 62 detects the hardware fault of the server including: the basic input and output system device judges whether the hardware interrupt signal of the server arrives, and if so, the The basic input and output system device detects the relevant hardware of the operating system; obtains the fault information of the hardware.

In this embodiment, it also includes a storage setting module 64, which is used to allocate a storage device for storing the hardware fault information on the server serial flash memory before the fault information storage module stores the fault information. fault storage area.

In the present invention, a server is also provided, and the server includes the above-mentioned basic input output system device.

The technical solution provided by the present invention can be widely applied to computers, network communication equipment and other equipment. The basic input and output system device can detect the faults of the hardware equipment in the whole cycle of the server operation, which can prevent the server from Failures occur, improving the stability and reliability of the server operation.

The above content is a further detailed description of the present invention in conjunction with specific embodiments, and it cannot be assumed that the specific implementation of the present invention is limited to these descriptions. For those of ordinary skill in the technical field of the present invention, without departing from the concept of the present invention, some simple deduction or replacement can be made, which should be regarded as belonging to the protection scope of the present invention.

Claims (10)

1. a kind of server hardware fault detection method, it is characterised in that include：

The basic input output system device of server detects the server and enters startup stage；

The basic input output system device starts to carry out hardware fault in each working stage to the server Detection, the working stage includes the startup stage；

The basic input output system device will detect that the hardware fault information for obtaining is stored.

2. server failure detection method as claimed in claim 1, it is characterised in that the startup Stage includes initial phase, and the basic input output system device is in the initial phase to the clothes Business device carries out hardware failure detection to be included：

The hardware detection mechanism that the basic input output system device is provided according to the server is to the clothes At least one of the business CPU of device, internal memory, chipset and power supply carry out the pre-detection of hardware and obtain current Hardware information, faulty hardware information is filtered out from the hardware information is analyzed process and obtain accordingly Hardware fault information.

3. server hardware fault detection method as claimed in claim 2, it is characterised in that described Startup stage also includes the device enumeration stage, and the basic input output system device is in the device enumeration rank Section carries out hardware failure detection to the server to be included：

The basic input output system device obtains the status information and resource letter of each hardware on the server Breath, and therefrom recognize the fault message of the hardware for breaking down.

4. the server hardware fault detection method as described in any one of claim 1-3, its feature exists In the startup stage is cold-start phase or thermal starting stage.

5. the server hardware fault detection method as described in any one of claim 1-3, its feature exists In, the working stage also include operating system pre-boot phase and in the operating system service operation stage extremely It is few one.

6. server hardware fault detection method as claimed in claim 5, it is characterised in that described When working stage includes operating system pre-boot phase, the basic input output system device is in the operation System pre-boot phase carries out hardware failure detection to the server to be included：

The basic input output system device is to the hardware device outside the server band that will be booted up Carry out pre-detection；

Obtain the Current hardware information of the hardware device；

The fault message of the hardware device for breaking down is filtered out from the Current hardware information；

When the working stage includes the operating system service operation stage, the basic input output system device Carrying out hardware failure detection to the server in the operating system service operation stage includes：It is described basic Input-output system device judges whether the hardware interrupt of the server arrives, if so, the then base This input-output system device is detected to the related hardware of the operating system；Obtain the event of the hardware Barrier information.

7. the server hardware fault detection method as described in any one of claim 1-3, its feature exists In before the basic input output system device will detect that the fault message that obtains is stored, also wrapping Include and distribute a failure for being used to store the hardware fault information on the server serial flash memorizer Memory block.

8. a kind of basic input output system device, it is characterised in that include：

Fault message detects trigger module, and whether startup stage is entered for detection service device；

Fault message detection module, enters for detecting the server in the fault detect trigger module During startup stage, start to carry out hardware failure detection, the work rank in each working stage to the server Section includes the startup stage；

Fault message memory module, for the hardware fault information that fault message detection module detection is obtained Stored.

9. basic input output system device as claimed in claim 8, it is characterised in that also include Storage setup module, for the hardware fault information to be carried out into storage in the fault message memory module Before, an event for being used to store the hardware fault information is distributed on the server serial flash memorizer Barrier memory block.

10. a kind of server, it is characterised in that including basic input as claimed in claim 8 or 9 Output system device.