CN112231165B - Memory fault simulation method based on link library injection and memory mapping modes - Google Patents

Abstract

The invention relates to a memory fault simulation method based on a link library injection and memory mapping mode, which comprises the following steps: the fault simulator is started on the target server, a command interface is initialized, a user inputs a fault simulation command, the fault simulator transmits the command to Command Dispatcher and Command Dispatcher to analyze the fault simulation command after receiving the fault simulation command, and a corresponding command processor is found according to a command mapping table; and the command processor completes corresponding fault injection according to the corresponding fault simulation command. The memory fault simulation method based on the link library injection and memory mapping modes can simulate the memory fault on the premise of not damaging the memory hardware, and is used for verifying the disaster tolerance of software to the memory fault. With non-intrusive transparency, the original hardware does not need to be destroyed and the original code does not need to be modified.

Description

Memory fault simulation method based on link library injection and memory mapping modes

Technical Field

The invention relates to the technical field of memory faults, in particular to a memory fault simulation method based on a link library injection and memory mapping mode.

Background

With the development of technology, computers have entered into thousands of households. The hardware system of the computer is composed of an arithmetic unit, a controller, a memory, an input device and an output device. The memory in a computer is divided into a memory and an external memory. The memory is used for storing programs and data which are currently used or are used at any time. Once the memory is in error or fault, the program can not work normally or be down. Therefore, the method has important significance in researching possible faults of the memory.

Memory failure is one of the most common failures in computer systems, and in critical systems with extremely high availability requirements, it is necessary to evaluate their fault tolerance to memory failures. The most effective method for evaluating fault tolerance mechanism by fault injection. The fault injection is to aim at specific fault models, consciously promote faults in a target system, accelerate the occurrence of errors and failures, and verify the fault tolerance, the fault safety and other information by analyzing the response information of the system to the injected faults.

Disclosure of Invention

The invention provides a memory fault simulator which simulates memory faults on the premise of not damaging memory hardware by means of link library injection and memory mapping, and is used for verifying the disaster tolerance of software to the memory faults.

The invention relates to a memory fault simulation method based on a link library injection and memory mapping mode, which comprises the following steps:

1. the memory fault simulation method based on the link library injection and memory mapping modes is characterized by comprising the following steps of:

A. the fault simulator is started on the target server, a command interface is initialized, and a user waits for inputting a fault simulation command;

B. the user inputs fault simulation commands, including a memory allocation failure command and a memory shortage command;

C. after receiving the fault simulation command, the fault simulator transmits the fault simulation command to Command Dispatcher and Command Dispatcher to analyze the fault simulation command, and finds a corresponding command processor according to the command mapping table;

D. and the command processor completes corresponding fault injection according to the corresponding fault simulation command.

The invention simulates the memory fault on the premise of not damaging the memory hardware, is used for verifying the disaster tolerance of software to the memory fault, has non-invasive transparency, does not need to damage the original hardware, and does not need to modify the original code.

Further, step D includes: the fault simulation command comprises a memory allocation failure command and a memory shortage command;

the memory allocation failure command is executed as follows:

after receiving the allocation failure command, the fault simulator sets an LD_PRELOAD environment variable to an operating system, loads newmalloc.so with a memory allocation failure function in advance, covers malloc and brk functions in glibc.so, and when software applies for a memory through the malloc and brk functions, memory allocation failure occurs;

the execution steps of the memory shortage command are as follows:

after receiving the memory shortage command, the command processor firstly executes mkdir-p/mnt/tmpfs & & mount tmpfs/mnt/tmpfs-t tmpfs to map the file into the memory; and then circularly executing dd if=/dev/zero of=/mnt/tmpfsbs=1m and then writing 1M data into the memory mapping file, wherein the memory of the system is continuously reduced, namely the memory deficiency occurs.

The command principle of the step D memory allocation failure is as follows: the software process runs on a run time provided by the operating system, which provides malloc, brk functions for software use to apply for memory to the operating system, and any memory allocation function is ultimately implemented using brk. The malloc, brk functions are located in glibc.

The command principle of the step D of insufficient memory is as follows: the operating system supports mapping a part of the memory into a file through memory mapping, and can conveniently consume a large amount of memory of the operating system by writing a large amount of data into the mapped file, so that the fault of insufficient memory is simulated.

The memory fault simulation method based on the link library injection and memory mapping modes can simulate the memory fault on the premise of not damaging the memory hardware, and is used for verifying the disaster tolerance of software to the memory fault. With non-intrusive transparency, the original hardware does not need to be destroyed and the original code does not need to be modified.

Drawings

FIG. 1 is a schematic diagram of a memory failure simulation architecture according to the present invention.

FIG. 2 is a schematic diagram of an architecture for simulating memory starvation caused by memory failure according to the present invention.

Detailed Description

The foregoing of the invention will be described in further detail with reference to the following detailed description of the examples. It should not be understood that the scope of the above subject matter of the present invention is limited to the following examples only. Various substitutions and alterations are also possible, without departing from the spirit of the invention, and are intended to be within the scope of the invention.

The architecture of the present invention as shown in figure 1 simulates the memory failure resulting in the application memory failure,

the fault simulator sets an LD_PRELOAD environment variable to an operating system, loads newmalloc.so with a memory allocation failure function in advance, covers malloc and brk functions in glibc.so, and when software applies for a memory through the malloc or brk functions, memory allocation failure occurs;

the fault simulator is started on the target server, a command interface is initialized, and a user inputs a memory allocation failure command; after receiving the memory allocation failure command, the fault simulator transmits the command to Command Dispatcher and Command Dispatcher to analyze the fault simulation command, and finds a corresponding command processor according to a command mapping table; after triggering a memory allocation failure command, the fault simulator sets an LD_PRELOAD environment variable to an operating system, loads newmalloc.so with a memory allocation failure function in advance, covers malloc and brk functions in glibc.so, and when software applies for a memory through the malloc or brk functions, memory allocation failure occurs;

as shown in FIG. 2, the architecture of the present invention simulates the memory failure resulting in insufficient memory

The fault simulator first executes mkdir-p/mnt/tmpfs & & mount tmpfs/mnt/tmpfs-t tmpfs to map the file into memory. Then, the dd if=/dev/zero of=/mnt/tmpfsbs=1m is circularly executed, and 1M data is written into the memory mapping file, at this time, the system memory is continuously reduced, and the memory deficiency simulation is successful.

The specific steps are that the fault simulator is started on the target server, a command interface is initialized, a user inputs a memory shortage command, the fault simulator transmits the command to Command Dispatcher and Command Dispatcher to analyze the fault simulation command after receiving the memory shortage command, and a corresponding command processor is found according to a command mapping table; after the command processor triggers the memory shortage command, the fault simulator firstly executes mkdir-p/mnt/tmpfs & & mount tmpfs/mnt/tmpfs-t tmpfs to map the file into the memory. Then, the dd if=/dev/zero of=/mnt/tmpfsbs=1m is circularly executed, and 1M data is written into the memory mapping file, at this time, the system memory is continuously reduced, and the memory deficiency simulation is successful.

Claims (1)

1. The memory fault simulation method based on the link library injection and memory mapping modes is characterized by comprising the following steps of:

A. the fault simulator is started on the target server, a command interface is initialized, and a user waits for inputting a fault simulation command;

B. the user inputs fault simulation commands, including a memory allocation failure command and a memory shortage command;

C. after receiving the fault simulation command, the fault simulator transmits the fault simulation command to Command Dispatcher and Command Dispatcher to analyze the fault simulation command, and finds a corresponding command processor according to the command mapping table;

D. the command processor triggers corresponding faults according to the corresponding fault simulation commands;

the step D comprises the following steps: the fault simulation command comprises a memory allocation failure command and a memory shortage command;

the memory allocation failure command is executed as follows:

after receiving the allocation failure command, the fault simulator sets an LD_PRELOAD environment variable to an operating system, loads newmalloc.so with a memory allocation failure function in advance, covers malloc and brk functions in glibc.so, and when software applies for a memory through the malloc and brk functions, memory allocation failure occurs;

the execution steps of the memory shortage command are as follows:

after receiving the memory shortage command, the command processor firstly executes mkdir-p/mnt/tmpfs & & mount tmpfs/mnt/tmpfs-t tmpfs to map the file into the memory; and then circularly executing dd if=/dev/zero of=/mnt/tmpfsbs=1m and then writing 1M data into the memory mapping file, wherein the memory of the system is continuously reduced, namely the memory deficiency occurs.

Images