CN107247505A - A kind of Cloud Server power supply blackbox design methods for being easy to check - Google Patents

Abstract

An easy-to-view blackbox design method for cloud server power supply, including the following steps: when the server power supply fails, classify the data at the time of the failure and store them in different registers, and the failure data is identified by page; in the same register The data category for storing the fault time is the same; when querying the server power failure information to troubleshoot the fault, switch to the corresponding page according to the time when the fault occurs, and then use the standard PMbus specification command to read the data of the fault time stored in different registers. Data, and then complete the investigation and confirmation of the cause of the server power failure based on the read and parsed data. The invention changes the storage of server power failure information data from the existing single register to multiple registers to solve the problem that the server command cannot be read when the blackbox data length is too long, and facilitates the management and operation of power failure.

Description

An easy-to-view design method of cloud server power supply blackbox

technical field

The invention relates to a method for fast data storage and reading, and belongs to the technical field of computers, in particular to an easy-to-view design method of a cloud server power supply blackbox.

Background technique

The server power supply refers to the power supply used on the server. It is a switching power supply just like the power supply used by the personal computer. According to the standard, the server power supply can be divided into two types: ATX power supply and SSI power supply. When the server power supply fails, in order to locate the fault problem, it is necessary to write the fault information inside the power supply for troubleshooting, that is, blackbox data, which is called black box fault log in Chinese.

For blackbox data recording, most of the existing solutions are to directly put multiple sets of data in a register, read out all the address bit information when it needs to be used, and then analyze the corresponding data according to the specific segment. The instruction to read power information inside the server has requirements for the length of the read. When the length of the blackbox design is greater than the requirements of the server itself, the instruction cannot read the information of the blackbox of the power supply. The power supply must be taken out of the server and used by the server manufacturer. Analysis is not convenient for on-site personnel to operate, and sometimes it cannot meet the customer's fault analysis requirements.

For example, Chinese patent (application number CN201611208065.0) discloses "a server monitoring system based on server architecture". The server power supply data acquisition module is used to obtain the temperature, power consumption, output current, output voltage, and status information of the power supply. Fault status information; the processor receives the data information sent by the fan data acquisition module, the hard disk data acquisition module, and the server power supply data acquisition module, stores the received data information in the data storage module, and displays the received data information on the display unit at the same time , when the received data information exceeds the threshold, a prompt message is sent on the display unit. In the daily use and operation and maintenance of the server, it is required to be able to monitor the operating parameters of the server in real time, realize the visualization of the server operating parameter information and server fault information, and monitor the internal components and modules of the server in real time to ensure the stable operation of the server. This invention helps to improve the monitoring and operation of the server by visualizing different parameter values of the power supply. However, when the blackbox length of the power failure is greater than the server itself, the problem that the command cannot read the blackbox information is not involved, which is not convenient for the operator. Analyze and judge the cause of power failure, and then take effective solutions.

Contents of the invention

The invention provides an easy-to-view design method for the power supply blackbox of a cloud server, which is used to solve the problem that the instruction cannot read the information of the power supply blackbox when the instruction read inside the server is shorter than the length of the blackbox design, and is convenient for on-site personnel to analyze the power failure And operation, so as to solve the problems existing in the use of server power supply in a targeted manner.

The present invention is achieved through the following technical solutions:

An easy-to-view blackbox design method for cloud server power supply, including server power supply and multiple registers, includes the following steps:

S1. When the power supply of the server fails, the data at the time of the failure is classified by category and stored in different registers. The failure data is identified by page encoding;

S2, according to the steps of S1, record the data at the time when the power supply of the server fails next time, and the data categories at the time of failure are stored in the same register are the same;

S2. When querying the server power supply fault information to troubleshoot the fault, switch to the corresponding page code according to the time when the fault occurs, and then use the standard PMbus specification command to read the data at the time of the fault stored in different registers, and then according to the read The data obtained and analyzed completes the investigation and confirmation of the cause of the server power failure.

As mentioned above, an easy-to-view cloud server power supply blackbox design method, there are 4 registers set in the step S1.

As mentioned above, an easy-to-view cloud server power supply blackbox design method, the data stored by category in the four register settings in step S1 are respectively the input voltage, input current, output voltage, and air inlet temperature corresponding to the fault moment value.

According to an easy-to-view blackbox design method for a cloud server power supply as described above, the length of the data stored in the register by category in the step S1 ranges from 1 to 15 bytes.

According to an easy-to-view blackbox design method for cloud server power supplies as described above, the range of fault data identified by page coding in the step S1 is between 5 and 15.

Compared with prior art, the advantage of the present invention is:

1. By changing the server power failure information data from the existing single register to multiple registers, it can effectively solve the problem that the server command cannot read the data when the blackbox data length is too long.

2. The present invention can be used in all server systems that support Pmbus. Without using any scripts or software, the fault data can be read directly through the ipmitool command for power fault analysis, which is convenient for R&D and customer service personnel to read power fault information, saving problem analysis time.

3. The present invention can form a modular design and product, which can be used in subsequent various servers to meet the needs of users for server power management functions and reduce product development costs.

Description of drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the description of the embodiments or the prior art.

Fig. 1 is the schematic diagram of existing server power failure data blackbox access;

Fig. 2 is a schematic diagram of blackbox access of server power failure data according to the present invention.

detailed description

In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments It is a part of embodiments of the present invention, but not all embodiments.

An easy-to-view blackbox design method for cloud server power supply, including server power supply and multiple registers, includes the following steps:

S1. When the power supply of the server fails, the data at the time of the failure is classified by category and stored in different registers. The failure data is identified by page encoding;

S2, according to the steps of S1, record the data at the time when the power supply of the server fails next time, and the data categories at the time of failure are stored in the same register are the same;

S2. When querying the server power supply fault information to troubleshoot the fault, switch to the corresponding page code according to the time when the fault occurs, and then use the standard PMbus specification command to read the data at the time of the fault stored in different registers, and then according to the read The data obtained and analyzed completes the investigation and confirmation of the cause of the server power failure.

Specifically, in this embodiment, the power supply blackbox information is stored in multiple independent registers such as Pin and Iin, and then the fault information is read from the above registers independently according to needs, and at the same time, in order to distinguish the information recorded each time, it is stored according to the page , so that it can be easily operated in any server without worrying that the length of the data instruction exceeds the requirements in the server.

Among them, the principle of blackbox information storage is as follows:

When a power failure occurs, the data at the time of the failure will be recorded, and placed in the corresponding page according to the time of occurrence, that is, all the data of the first failure will be recorded in page1, and the second time will be recorded in page2, which will be incremented to page one by one. The maximum range, and then loop; when reading, switch to the corresponding page through the command, and then read the data according to the standard PMbus specification command.

For further clarity, the record of event 1 is taken as an example for illustration, and event 1 represents a power failure of one of the servers.

As shown in Figure 1, the existing server power failure data blackbox access process is as follows:

①. When the power supply fails in event 1, the power supply will record the data A to data N to be recorded in a register in a certain order;

②. When it is necessary to obtain event 1 failure from inside the power supply, it is necessary to read the data of the entire register. Generally, the server power failure data corresponding to event 1 is several hundred bytes, and then intercept data A to data N in event 1. data, and analyze them separately according to the bytes occupied by each data;

③. The above operations are generally applicable to the scripts developed by the manufacturers themselves, but they cannot be used in the server if the byte length is too long.

As shown in Figure 2, the improved server power failure data blackbox access process of the present invention is as follows:

①. When the power supply fails in event 1, the power supply will record the event data that needs to be recorded in the registers corresponding to data A to data N;

②. When it is necessary to obtain the event 1 fault from inside the power supply, it is only necessary to switch to page1 through the command, and then use the standard ipmitool command to read the data in page1 from data A to data N;

③. When you need to view event N, you need to switch to page N, and then read it according to the method in ②.

It should be noted that N is limited by the storage space and is determined according to the storage space, generally between 5 and 15. The above operations are not only suitable for reading data by the power supplier itself, but also suitable for reading in any server system that supports Pmbus, and the operation is more convenient.

In conjunction with Fig. 2, the specific implementation steps of the present invention are as follows:

① When the power supply fails once, the data at the time of the failure, including input voltage, input current, output voltage, air inlet temperature and other information, will be recorded in multiple registers in the form of an event;

②. When you need to troubleshoot the fault based on the data recorded in event 1, you need to switch the power to page1 first. page is a virtual page representing all the data in the event. In fact, it is the first or Nth record of each register from data A to data N. A collection of data, so that when viewing event 1, the data of other events will not be read;

③. Use the ipmitool command to read the input voltage, air inlet temperature, output voltage, status word and other data information in page1 respectively. The length of each type of data in a single register is only 1-2, and the maximum is no more than 15 bytes of data;

④. According to the read and analyzed data, judge the specific fault and possible fault cause of event 1.

According to the above steps, switch to different page codes and read different event (1 to N) information to complete the troubleshooting and confirmation of power failures. In addition to the above data, the total power-on time of the power supply can also be recorded according to actual needs to judge the occurrence of faults.

The technical contents not described in detail in the present invention are all known technologies.

Claims (5)

1. a kind of Cloud Server power supply blackbox design methods for being easy to check, including server power supply and multiple deposits Device, it is characterised in that comprise the following steps：

S1, when server power supply breaks down, the data category of fault moment is distinguished, and stores to different registers In, this time fault data is identified with page coded systems；

S2, break down the data at moment according to recording next server power supply the step of S1, failure is deposited in same register The data category at moment is identical；

S2, when inquire about fault of server power supply information to investigate failure when, at the time of generation according to failure instruction be switched to correspondence Page codings, the PMbus standardized instructions for reusing standard read the number for being respectively stored in fault moment in different registers According to then according to the investigation for the data completion fault of server power supply reason for reading and being resolved to and confirmation.

2. a kind of Cloud Server power supply blackbox design methods for being easy to check according to claim 1, its feature exists In register is provided with 4 in the step S1.

3. a kind of Cloud Server power supply blackbox design methods for being easy to check according to claim 2, its feature exists In 4 registers set the data of middle category storage in the step S1, respectively the corresponding input voltage of fault moment, defeated Enter electric current, output voltage, air intake vent Temperature numerical.

4. according to a kind of any described Cloud Server power supply blackbox design methods for being easy to check of claims 1 to 3, its It is characterised by, category is stored in the data length scope of register between 1~15 bytes in the step S1.

5. a kind of Cloud Server power supply blackbox design methods for being easy to check according to claim 1, its feature exists In the scope that fault data is identified by page modes in the step S1 is between 5~15.