Method, device, electronic equipment and storage medium for continuous availability of cloud host

Abstract

The application relates to a cloud host continuous availability method, device, electronic equipment and storage medium. The cloud host continuous availability method comprises the following steps: S1, a sampling test module acquires a server memory CE value; S2, a sampling collection module is responsible for periodically collecting the memory information sampled by the sampling test module, recording the server memory CE value, and comparing the server memory CE value with a preset threshold value; S3, a pre-warning reminding module judges a memory pre-fault according to a preset CE value judgment logic, sends an alarm when a pre-warning reminding threshold value is triggered, and triggers a migration management module to send a cloud host hot migration instruction; and S4, the migration management module receives the alarm of the pre-warning reminding module, triggers a fault processing threshold value, and sends a migration management command to perform virtual machine migration. According to the cloud host continuous availability method, system downtime caused by real memory faults can be avoided in advance, and the continuous availability of business is ensured.

Description

Technical Field

[0001] This invention relates to the field of IT and software development, and more specifically, to a method, apparatus, electronic device, and storage medium for the continuous availability of cloud hosts based on EDAC. Background Technology

[0002] EDAC (Error Detection and Correction) is the current operating system's error correction and detection framework. Its main purpose is to enable the operating system to detect and report hardware errors during operation.

[0003] CE (correctable error) refers to a correctable error. The server will repair erroneous data in memory. The health status of memory is usually judged by the size of the CE value and the reference value.

[0004] With the increasing number of enterprises using public clouds, the scale of the underlying host machines of public cloud service providers is growing larger and larger. Consequently, the failure rate of these underlying host machines is increasing exponentially. Statistics show that the top three hardware components with the highest failure rates in server hardware are: hard drives, power supplies, and memory. While hard drives are supported by RAID array technology for redundancy, and hard drive failures generally do not affect the running host machine, host power supplies also offer redundancy. Memory, however, is a high-failure-rate component without redundancy. If memory fails, the system will experience various anomalies such as system crashes, host failures, and blue screens, immediately halting services on the host machine. If a host machine running cloud servers experiences a memory failure, the cloud servers on the failed host will be unable to run their operating systems properly. Severe memory failures may cause the host machine to restart, requiring a cold migration of the cloud servers to other healthy host machines in the cluster, resulting in service interruption and affecting continuous service availability. Most existing methods for obtaining host memory fault alarms rely on the hardware status records of the server's BMC. When an alarm is triggered in the BMC, the memory is already unusable or faulty, making fault handling very passive.

[0005] Therefore, how to detect memory failures early has become an urgent problem to be solved. Summary of the Invention

[0006] The technical problem this invention aims to solve is how to detect host machine memory failures early, provide early warnings before memory becomes unusable, and automatically migrate cloud hosts on the host machine to normal host machines in the cluster if the memory is identified as a pre-failure memory, ensuring continuous service availability.

[0007] To address the aforementioned technical problems, according to one aspect of the present invention, a method for ensuring the continuous availability of a cloud server is provided. The method is implemented using a device for ensuring the continuous availability of a cloud server, the device comprising a sampling and testing module, a sampling and collection module, an early warning and alert module, and a migration management module. The method for ensuring the continuous availability of a cloud server includes the following steps: S1. The sampling and testing module acquires the CE value of the server memory. The sampling and testing module is responsible for periodically acquiring system memory information and corresponding host memory slot information on the host machine, periodically acquiring the CE value corresponding to each memory module, and matching the basic memory information with the CE value information; S2. The sampling and collection module is responsible for periodically collecting the memory information sampled by the sampling and testing module, recording and processing the server memory CE value, and comparing it with a pre-set threshold; S3. The early warning and alert module, according to a preset CE value judgment logic, judges based on two dimensions: CE difference comparison and cumulative CE value summation. When the CE difference or cumulative CE value summation reaches the triggering early warning and alert threshold, an alarm is sent, indicating a pre-fault in the memory, and triggering the migration management module to send a cloud server hot migration command; S4. Upon receiving the alarm from the early warning and alert module, the migration management module triggers a fault handling threshold and sends a migration management command to migrate the virtual machine.

[0008] According to an embodiment of the present invention, the sampling and collection module sets a first threshold and a second threshold for CE difference, wherein the first threshold for CE difference is less than the second threshold for CE difference. The sampling and collection module is responsible for periodically calculating the CE difference. The sampling and collection module compares the CE difference collected each time with the set first threshold and second threshold for CE difference. If the obtained CE difference is less than the first threshold for CE difference, it means that the change in the obtained CE difference is within the normal range. If the obtained CE difference is greater than the first threshold for CE difference but less than the second threshold for CE difference, it means that the change in the obtained CE difference exceeds the normal range, and it is added to the observation list. If the obtained CE difference is greater than the second threshold for CE difference, it is judged as a memory pre-fault, the host IP address and the pre-fault memory slot information are recorded, and the migration management module is triggered to perform a cloud host hot migration operation.

[0009] Furthermore, the CE value of memory of the same brand or batch was observed and verified when the memory failed, and the fault CE difference was determined. The first threshold of the CE difference was 50-70% of the fault CE difference, preferably 55-65%, more preferably 60%; the second threshold of the CE difference was 75-90% of the fault CE difference, preferably 78-95%, more preferably 80%.

[0010] According to an embodiment of the present invention, a first threshold and a second threshold for the total CE value are set in the sampling and collection module. The first threshold for the total CE value is less than the second threshold. The sampling and collection module is responsible for periodically calculating the total CE value. The sampling and collection module compares the obtained total CE value for each calculation with the set first threshold, second threshold, and third threshold. If the obtained total CE value is less than the first threshold, it indicates that the change in the obtained total CE value is within the normal range. If the obtained total CE value is greater than the first threshold but less than the second threshold, it indicates that the change in the obtained total CE value exceeds the normal range and is added to the observation list. If the obtained total CE value is greater than the second threshold, it is judged as a memory pre-fault, the host IP address and the pre-fault memory slot information are recorded, and the migration management module is triggered to perform a cloud host hot migration operation.

[0011] Furthermore, the CE values ​​of memory failures using the same brand or batch are observed and verified to determine the total CE value of the failure. The first threshold for the total CE value is 50-70% of the total CE value of the failure, preferably 55-65%, and more preferably 60%. The second threshold for the total CE value is 75-90% of the total CE value of the failure, preferably 78-95%, and more preferably 80%.

[0012] According to an embodiment of the present invention, step S4 includes the following steps: S41, for the hot migration operation of cloud hosts by the migration management module, firstly, obtain the IP address of the host machine whose memory CE value is greater than the second threshold; S42, obtain the list of all cloud hosts running on the host machine through the host machine IP address; S43, send instructions to the pre-fault memory virtualization cluster through the migration management module to perform hot migration of cloud hosts one by one to the normal host machine in the cluster; S44, check whether the cloud host migration task is successful. If successful, continue to the next one until all hot migrations are completed. If the task fails, send an alarm and perform manual intervention afterwards.

[0013] Furthermore, step S4 also includes:

[0014] S45. Report the recorded pre-memory fault host machine information for memory repair, and perform memory fault repair after the cloud host migration is completed.

[0015] According to a second aspect of the present invention, an apparatus for the continuous availability of a cloud host is provided, comprising: a sampling test module, which acquires server memory CE values, and is responsible for periodically acquiring system memory information and corresponding host memory slot information on the host machine, periodically acquiring the CE value corresponding to each memory stick, and matching the memory basic information with the CE value information; a sampling collection module, which is responsible for periodically collecting the memory information sampled by the sampling test module, recording and processing the server memory CE values, and comparing them with a preset threshold; an early warning module, which judges according to a preset CE value judgment logic, the judgment logic judging through two dimensions: CE difference comparison and cumulative CE value summation, and sends an alarm when the CE difference or cumulative CE value summation reaches the triggering early warning threshold, judging the memory as a pre-fault, and triggering the migration management module to send a cloud host hot migration command; and a migration management module, which, upon receiving the alarm from the early warning module, triggers a fault handling threshold and sends a migration management command to migrate the virtual machine. The sampling and collection module sets a first threshold and a second threshold for CE difference, where the first threshold is less than the second threshold. The module is responsible for periodically calculating the CE difference. It compares each collected CE difference with the set first and second thresholds. If the obtained CE difference is less than the first threshold, the change is within the normal range. If the obtained CE difference is greater than the first threshold but less than the second threshold, the change exceeds the normal range and is added to the observation list. If the obtained CE difference is greater than the second threshold, it is judged as a memory pre-fault, and the host IP address and pre-fault memory slot information are recorded, triggering the migration management module to perform a cloud host hot migration operation. The sampling and collection module sets a first threshold and a second threshold for the total CE value. The first threshold is less than the second threshold. The sampling and collection module is responsible for periodically calculating the total CE value. It compares the calculated total CE value with the set first and second thresholds. If the total CE value is less than the first threshold, the change in the total CE value is within the normal range. If the total CE value is greater than the first threshold but less than the second threshold, the change in the total CE value exceeds the normal range and is added to the observation list. If the total CE value is greater than the second threshold, it is judged as a memory pre-fault, the host IP address and the pre-fault memory slot information are recorded, and the migration management module is triggered to perform a cloud host hot migration operation.

[0016] According to a third aspect of the present invention, an electronic device is provided, comprising: a memory, a processor, and a cloud host continuous availability program stored in the memory and executable on the processor, wherein the cloud host continuous availability program, when executed by the processor, implements the steps of the cloud host continuous availability method described above.

[0017] According to a fourth aspect of the present invention, a computer storage medium is provided, wherein a cloud host persistent availability program is stored on the computer storage medium, and when the cloud host persistent availability program is executed by a processor, it implements the steps of the cloud host persistent availability method described above.

[0018] Compared with the prior art, the technical solution provided by the embodiments of the present invention can achieve at least the following beneficial effects:

[0019] This invention detects and acquires the host machine's memory operating status, using memory CE values ​​to determine the likelihood of an impending failure. This proactively avoids system crashes, host machine restarts, and other business-impacting anomalies that would occur when memory failures actually occur. Combined with virtualization hot migration technology, it allows cloud hosts to be migrated out of the pre-failure host machine without interruption.

[0020] This invention can avoid most host machine downtime failure scenarios, has wide applicability, and is easy to deploy. For businesses that require continuous availability, such as the financial industry and public service industry, it can greatly avoid major failures, improve data security, system stability, and ensure maximum business availability. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings of the embodiments will be briefly described below. Obviously, the drawings described below only relate to some embodiments of the present invention and are not intended to limit the present invention.

[0022] Figure 1 This is a flowchart illustrating a method for ensuring the continuous availability of a cloud server according to an embodiment of the present invention;

[0023] Figure 2 This is a diagram illustrating the warning and alert judgment logic according to an embodiment of the present invention. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the described embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] Unless otherwise defined, the technical or scientific terms used herein shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains. The terms “first,” “second,” and similar terms used in the specification and claims of this patent application do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Similarly, the terms “an” or “a” and similar terms do not indicate a limitation of quantity, but rather indicate the presence of at least one.

[0026] Figure 1 This is a flowchart illustrating a method for ensuring the continuous availability of a cloud server according to an embodiment of the present invention.

[0027] like Figure 1 As shown, a method for ensuring the continuous availability of cloud servers is implemented based on a device for ensuring the continuous availability of cloud servers. The device includes a sampling test module, a sampling collection module, an early warning and reminder module, and a migration management module.

[0028] The methods to ensure the continued availability of cloud servers include the following steps:

[0029] S1. The sampling test module obtains the CE value of the server memory. The sampling test module is responsible for periodically obtaining the system memory information and corresponding host memory slot information on the host machine, periodically obtaining the CE value corresponding to each memory stick, and matching the basic memory information with the CE value information.

[0030] S2. The sampling and collection module is responsible for periodically collecting and recording the memory information sampled by the sampling and testing module, processing the server memory CE value, and comparing it with a preset threshold.

[0031] S3. The early warning and reminder module judges according to the preset CE value judgment logic. The judgment logic judges based on two dimensions: CE difference comparison and cumulative CE value sum. When the CE difference or cumulative CE value sum reaches the triggering early warning and reminder threshold, an alarm is sent. If the memory is judged to be in a pre-fault state, the migration management module is triggered to send a cloud host hot migration command.

[0032] S4. When the migration management module receives an alarm from the early warning module, it triggers the fault handling threshold and sends a migration management command to migrate the virtual machine.

[0033] This invention detects and acquires the host machine's memory operating status, using memory CE values ​​to determine the likelihood of an impending failure. This proactively avoids system crashes, host machine restarts, and other business-impacting anomalies that would occur when memory failures actually occur. Combined with virtualization hot migration technology, it allows cloud hosts to be migrated out of the pre-failure host machine without interruption.

[0034] Figure 2 This is a diagram illustrating the warning and alert judgment logic according to an embodiment of the present invention.

[0035] like Figure 2 As shown, the sampling and collection module sets a first threshold and a second threshold for CE difference, wherein the first threshold is less than the second threshold. The sampling and collection module is responsible for periodically calculating the CE difference, and compares the collected CE difference with the set first and second thresholds for CE difference each time.

[0036] If the obtained CE difference is less than the first threshold of CE difference, it means that the change in the obtained CE difference is within the normal range; if the obtained CE difference is greater than the first threshold of CE difference but less than the second threshold of CE difference, it means that the change in the obtained CE difference exceeds the normal range, and it is added to the observation list; if the obtained CE difference is greater than the second threshold of CE difference, it is judged as a memory pre-fault, the host IP address and the pre-fault memory slot information are recorded, and the migration management module is triggered to perform a cloud host hot migration operation.

[0037] Furthermore, the CE value of memory of the same brand or batch was observed and verified when the memory failed, and the fault CE difference was determined. The first threshold of the CE difference was 50-70% of the fault CE difference, preferably 55-65%, more preferably 60%; the second threshold of the CE difference was 75-90% of the fault CE difference, preferably 78-95%, more preferably 80%.

[0038] According to one or more embodiments of the present invention, a first threshold and a second threshold for the total CE value are set in the sampling and collection module, wherein the first threshold for the total CE value is less than the second threshold for the total CE value, and wherein the sampling and collection module is responsible for periodically calculating the total CE value, and the sampling and collection module compares the obtained total CE value for each calculation with the set first threshold for the total CE value, the second threshold for the total CE value, and the third threshold for the total CE value.

[0039] If the total CE value is less than the first threshold, it means the change in the total CE value is within the normal range; if the total CE value is greater than the first threshold but less than the second threshold, it means the change in the total CE value exceeds the normal range and is added to the observation list; if the total CE value is greater than the second threshold, it is judged as a memory pre-failure, the host IP address and pre-failure memory slot information are recorded, and the migration management module is triggered to perform a cloud host hot migration operation.

[0040] Furthermore, the CE values ​​of memory failures using the same brand or batch are observed and verified to determine the total CE value of the failure. The first threshold for the total CE value is 50-70% of the total CE value of the failure, preferably 55-65%, and more preferably 60%. The second threshold for the total CE value is 75-90% of the total CE value of the failure, preferably 78-95%, and more preferably 80%.

[0041] According to one or more embodiments of the present invention, step S4 includes the following steps:

[0042] S41. For hot migration of cloud hosts in the migration management module, first obtain the host IP address whose memory CE value is greater than the second threshold.

[0043] S42. Obtain a list of all cloud hosts running on the host machine by using the host machine's IP address.

[0044] S43. Send instructions to the pre-fault memory virtualization cluster through the migration management module to perform hot migration of cloud hosts one by one to the normal host machine of the cluster.

[0045] S44. Check if the cloud server migration task is successful. If successful, continue to the next server until all hot migrations are completed. If the task fails, send an alarm and require manual intervention.

[0046] S45. Report the recorded pre-memory fault host machine information for memory repair, and perform memory fault repair after the cloud host migration is completed.

[0047] According to a second aspect of the present invention, an apparatus for the continuous availability of cloud hosts is provided, comprising: a sampling test module, a sampling collection module, an early warning and reminder module, and a migration management module.

[0048] The sampling test module obtains the server memory CE value. The sampling test module is responsible for periodically obtaining the system memory information and corresponding host memory slot information on the host machine, periodically obtaining the CE value corresponding to each memory stick, and matching the basic memory information with the CE value information.

[0049] The sampling and collection module is responsible for periodically collecting and recording the memory information sampled by the sampling and testing module, processing the server memory CE value, and comparing it with a preset threshold.

[0050] The early warning and reminder module judges according to the preset CE value judgment logic. The judgment logic judges based on two dimensions: CE difference comparison and cumulative CE value summation. When the CE difference or cumulative CE value summation reaches the triggering early warning and reminder threshold, an alarm is sent. If the memory is judged to be in a pre-fault state, the migration management module is triggered to send a cloud host hot migration command.

[0051] The migration management module receives an alarm from the early warning module, triggers the fault handling threshold, and sends a migration management command to migrate the virtual machine.

[0052] The sampling and collection module sets a first threshold and a second threshold for CE difference, where the first threshold is less than the second threshold. The module is responsible for periodically calculating the CE difference. It compares each collected CE difference with the set first and second thresholds. If the obtained CE difference is less than the first threshold, the change is within the normal range. If the obtained CE difference is greater than the first threshold but less than the second threshold, the change exceeds the normal range and is added to the observation list. If the obtained CE difference is greater than the second threshold, it is judged as a memory pre-fault, and the host IP address and pre-fault memory slot information are recorded, triggering the migration management module to perform a cloud host hot migration operation.

[0053] The sampling and collection module sets a first threshold and a second threshold for the total CE value. The first threshold is less than the second threshold. The sampling and collection module is responsible for periodically calculating the total CE value. It compares the calculated total CE value with the set first and second thresholds. If the total CE value is less than the first threshold, the change in the total CE value is within the normal range. If the total CE value is greater than the first threshold but less than the second threshold, the change in the total CE value exceeds the normal range and is added to the observation list. If the total CE value is greater than the second threshold, it is judged as a memory pre-fault, the host IP address and the pre-fault memory slot information are recorded, and the migration management module is triggered to perform a cloud host hot migration operation.

[0054] This invention can avoid most host machine downtime failure scenarios, has wide applicability, and is easy to deploy. For businesses that require continuous availability, such as the financial industry and public service industry, it can greatly avoid major failures, improve data security, system stability, and ensure maximum business availability.

[0055] In practice, host machine memory sampling obtains specific memory CE value parameters for each host machine through the Linux EDAC framework tools. Each host machine needs to have the edac-utils tool deployed, and the host machine's CPU hardware architecture must be confirmed, with matching driver modules installed.

[0056] The sampling test module can simulate erroneous data writing for simulation testing.

[0057] The sampling test module collects information using a general-purpose monitoring agent, filtering by regular expressions to obtain memory CE values. The module periodically acquires system memory information and corresponding memory slot information from the host machine. It also periodically acquires the CE value for each memory module and matches the basic memory information with the CE values.

[0058] The memory hardware status can be determined from two dimensions: the difference between the previous and subsequent CE values, and the cumulative sum of the CE values ​​for each memory module.

[0059] The sampling and collection module is responsible for periodically collecting CE difference calculations. By configuring relevant strategies in the sampling and collection module, corresponding policy operations can be performed based on the magnitude of the difference in the memory CE values ​​after they are obtained.

[0060] The sampling and collection module sets a first and second threshold for the CE difference. The threshold settings must be adjusted according to the actual hardware conditions; there are no absolute, universally applicable values, as the hardware quality and characteristics of each manufacturer vary. Initially, threshold baseline statistics and data collection can be performed to observe and verify the CE value when using memory from the same brand or batch that fails. This value can then be used as a benchmark to set the second threshold, typically 80% of the fault CE value. The first threshold is then 60% of the fault CE value. After adjusting the thresholds to suit your hardware, you can then set the first and second thresholds.

[0061] After obtaining the CE value, the CE difference calculated each time is compared. If the obtained CE difference is less than the first threshold of CE difference, it means that the change of the obtained CE difference is within the normal range.

[0062] If the acquired CE difference is greater than the first threshold but less than the second threshold, it indicates that the change in the acquired CE difference exceeds the normal range. In this case, the value is added to the observation list, and an alarm is issued. Alarm methods can be configured via SMS, email, and other means.

[0063] If the obtained CE difference value is greater than the second threshold of CE difference value, it indicates a memory pre-failure. The host machine IP address and the pre-failure memory slot information are recorded, and an alarm message is issued. The alarm method is as described above and can also be customized. Furthermore, a command to perform a cloud host hot migration operation is sent in the migration management module. The operation command can be performed by calling the virtualization cluster API interface.

[0064] The sampling and collection module is responsible for periodically collecting and calculating the sum of CE values. By configuring relevant strategies in the sampling and collection module, corresponding policy operations can be performed on the sum of the obtained in-memory CE values.

[0065] The sampling and collection module sets a first threshold and a second threshold for the sum of CE values. The sum of CE values ​​obtained in each calculation is compared; if the sum of CE values ​​is less than the first threshold, it indicates that the change in the sum of CE values ​​is within the normal range.

[0066] If the total CE value is greater than the first threshold but less than the second threshold, it indicates that the change in the total CE value exceeds the normal range. In this case, the value will be added to the observation list, and an alarm will be issued. The alarm can be configured via SMS, email, or other methods.

[0067] If the sum of the acquired CE values ​​exceeds the second threshold, it indicates a memory pre-failure. The host machine IP address and the pre-failure memory slot information are recorded, and an alarm is issued. The alarm method is as described above and can also be customized. Furthermore, a command to perform a cloud host hot migration operation is sent in the migration management module. This operation command can be performed by calling the virtualization cluster API interface.

[0068] For the above-mentioned migration management module's hot migration operation of cloud hosts, the host machine IP address that is greater than the second threshold is first obtained.

[0069] Obtain a list of all cloud hosts running on the host machine by using the host machine's IP address.

[0070] The migration management module sends instructions to the virtualization cluster where the pre-failed host is located via API calls to perform hot migration of cloud hosts one by one.

[0071] Check if the cloud server migration task is successful. If successful, proceed to the next server until all hot migrations are complete. If the task fails, send an alert and require manual intervention.

[0072] The recorded host information of the pre-memory fault will be reported for memory repair. After the cloud host migration is completed, the memory fault will be repaired.

[0073] According to another aspect of the present invention, a device for continuous cloud server availability is provided, comprising: a memory, a processor, and a cloud server continuous availability program stored in the memory and executable on the processor, wherein the cloud server continuous availability program implements the steps of the cloud server continuous availability method described above when executed by the processor.

[0074] The present invention also provides a computer storage medium.

[0075] The computer storage medium stores a cloud server persistent availability program, which, when executed by the processor, implements the steps of the cloud server persistent availability method described above.

[0076] The method implemented when the cloud host continuous availability program running on the processor is executed can be referred to in various embodiments of the cloud host continuous availability method of the present invention, and will not be repeated here.

[0077] The present invention also provides a computer program product.

[0078] The computer program product of the present invention includes a cloud host continuous availability program, which, when executed by a processor, implements the steps of the cloud host continuous availability method as described above.

[0079] The method implemented when the cloud host continuous availability program running on the processor is executed can be referred to in various embodiments of the cloud host continuous availability method of the present invention, and will not be repeated here.

[0080] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0081] The above description is merely an exemplary embodiment of the present invention and is not intended to limit the scope of protection of the present invention, which is determined by the appended claims.

Claims

1. A method for ensuring the continuous availability of a cloud server, the method being implemented based on a device for ensuring the continuous availability of a cloud server, the device comprising a sampling and testing module, a sampling and collection module, an early warning and reminder module, and a migration management module. The method for ensuring the continued availability of the cloud server includes the following steps: S1. The sampling test module obtains the CE value of the server memory. The sampling test module is responsible for periodically obtaining the system memory information and corresponding host memory slot information on the host machine, periodically obtaining the CE value corresponding to each memory, and matching the basic memory information with the CE value information. S2. The sampling and collection module is responsible for periodically collecting the memory information sampled by the sampling and testing module and recording the server memory CE value, and comparing it with a preset threshold. S3. The early warning module judges according to the preset CE value judgment logic. The judgment logic judges by two dimensions: CE difference comparison and cumulative total CE value. When the CE difference or cumulative total CE value reaches the triggering early warning threshold, an alarm is sent. The memory is judged to be in pre-fault condition, and the migration management module is triggered to send a cloud host hot migration command. S4. Upon receiving an alarm from the early warning module, the migration management module triggers the fault handling threshold and sends a migration management command to migrate the virtual machine. The sampling and collection module sets a first threshold for CE difference and a second threshold for CE difference, wherein the first threshold for CE difference is less than the second threshold for CE difference. The sampling and collection module is responsible for periodically calculating the CE difference. The module compares each collected CE difference with a set first threshold and a second threshold for the CE difference. Specifically, if the obtained CE difference is less than the first threshold, it indicates that the change in the obtained CE difference is within the normal range; if the obtained CE difference is greater than the first threshold but less than the second threshold, it indicates that the change in the obtained CE difference exceeds the normal range, and it is added to the observation list; if the obtained CE difference is greater than the second threshold, it is judged as a memory pre-fault, the host machine IP address and the pre-fault memory slot information are recorded, and the migration management module is triggered to perform a cloud host hot migration operation. Among them, the CE value of memory of the same brand or batch when it fails is observed and verified to determine the fault CE difference. The first threshold of the CE difference is 50~70% of the fault CE difference; the second threshold of the CE difference is 75~90% of the fault CE difference.

2. The method as described in claim 1, wherein, Observe and verify the CE value when using memory of the same brand or batch that fails, and determine the fault CE difference. The first threshold of the CE difference is 55~65% of the fault CE difference; the second threshold of the CE difference is 78~95% of the fault CE difference.

3. The method as described in claim 1, wherein, In the sampling and collection module, a first threshold for total CE and a second threshold for total CE are set, wherein the first threshold for total CE is less than the second threshold for total CE. The sampling and collection module is responsible for periodically calculating the sum of CE values. This module compares the calculated sum of CE values ​​with a set first threshold for the sum of CE values ​​and a second threshold for the total CE value. Specifically, if the total CE value is less than the first threshold, it means that the change in the total CE value is within the normal range; if the total CE value is greater than the first threshold but less than the second threshold, it means that the change in the total CE value exceeds the normal range and is added to the observation list; if the total CE value is greater than the second threshold, it is judged as a memory pre-failure, the host IP address and the pre-failure memory slot information are recorded, and the migration management module is triggered to perform a cloud host hot migration operation.

4. The method of claim 3, wherein, Observe and verify the CE value when using memory of the same brand or batch that fails, and determine the total CE value of the failure. The first threshold of the total CE value is 50-70% of the total CE value of the failure; the second threshold of the total CE value is 75-90% of the total CE value of the failure.

5. The method of claim 1, wherein, Step S4 includes the following steps: S41. For hot migration of cloud hosts in the migration management module, first obtain the host machine IP address whose memory CE value is greater than the second threshold. S42. Obtain a list of all cloud hosts running on the host machine using the host machine's IP address; S43. Send instructions to the pre-fault memory virtualization cluster through the migration management module to perform hot migration of cloud hosts one by one to the normal host machine in the cluster. S44. Check if the cloud server migration task is successful. If successful, continue to the next server until all hot migrations are completed. If the task fails, send an alarm and then manually intervene.

6. The method of claim 5, wherein, Step S4 also includes: S45. Report the recorded pre-memory fault host machine information for memory repair, and perform memory fault repair after the cloud host migration is completed.

7. A device for ensuring the continuous availability of a cloud server, comprising: The sampling test module acquires the CE value of the server memory. The sampling test module is responsible for periodically acquiring the system memory information and corresponding host memory slot information on the host machine, periodically acquiring the CE value corresponding to each memory stick, and matching the basic memory information with the CE value information. The sampling and collection module is responsible for periodically collecting the memory information sampled by the sampling and testing module, recording the server memory CE value, and comparing it with a preset threshold. The early warning and reminder module judges according to a preset CE value judgment logic. The judgment logic judges based on two dimensions: CE difference comparison and cumulative CE value summation. When the CE difference or cumulative CE value summation reaches the triggering early warning and reminder threshold, an alarm is sent. The memory is judged to be in a pre-fault state, and the migration management module is triggered to send a cloud host hot migration command. The migration management module receives an alarm from the early warning module, triggers a fault handling threshold, and sends a migration management command to migrate the virtual machine. The sampling and collection module sets a first threshold and a second threshold for CE difference, wherein the first threshold is less than the second threshold. The module is responsible for periodically calculating the CE difference. It compares each collected CE difference with the set first and second thresholds. If the obtained CE difference is less than the first threshold, the change is within the normal range. If the obtained CE difference is greater than the first threshold but less than the second threshold, the change exceeds the normal range and is added to the observation list. If the obtained CE difference is greater than the second threshold, it is judged as a memory pre-fault, the host IP address and pre-fault memory slot information are recorded, and the migration management module is triggered to perform a cloud host hot migration operation. The sampling and collection module sets a first threshold and a second threshold for the total CE value. The first threshold is less than the second threshold. The sampling and collection module is responsible for periodically calculating the total CE value. The module compares the calculated total CE value with the set first and second thresholds. If the total CE value is less than the first threshold, the change in the total CE value is within the normal range. If the total CE value is greater than the first threshold but less than the second threshold, the change in the total CE value exceeds the normal range and is added to the observation list. If the total CE value is greater than the second threshold, it is judged as a memory pre-fault, the host IP address and the pre-fault memory slot information are recorded, and the migration management module is triggered to perform a cloud host hot migration operation.

8. An electronic device, comprising: The cloud server persistent availability program, stored on the memory and executable on the processor, wherein when executed by the processor, the cloud server persistent availability program implements the steps of the cloud server persistent availability method as described in any one of claims 1 to 6.

9. A computer storage medium, wherein, The computer storage medium stores a cloud host continuous availability program, which, when executed by a processor, implements the steps of the cloud host continuous availability method as described in any one of claims 1 to 6.

Images