Method for controlling large-scale upgrading

Abstract

The invention relates to the technical field of equipment upgrade management, and discloses a method for controlling large-scale upgrade. According to the method, after an upgrade package is verified, a hardware evaluation coefficient of each piece of equipment is obtained, after the integrity verification of the upgrade package is passed, an upgrade stability coefficient of each piece of equipment is constructed, and the upgrade stability coefficient corresponding to the hardware evaluation coefficient of each piece of equipment is synthesized to construct an overall upgrade stability coefficient. Judging whether a stability adjustment strategy needs to be executed or not, then reading a historical busy sequence table of the equipment, comprehensively calculating business busy evaluation coefficients of all the equipment in each time period, determining upgrading starting time, and after upgrading of any batch of equipment is completed, performing matching analysis on the currently upgraded equipment; according to the method, the equipment in the next batch is analyzed, whether a backspacing mechanism is triggered or not is judged, and after the backspacing mechanism is triggered, the next batch is continuously analyzed until all the equipment is upgraded, so that one-time operation of large-scale upgrading management is realized, and the controllability of the upgrading process is ensured.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of device upgrade management, in particular to a method for controlling large-scale upgrade. BACKGROUND

[0002] With the wide application of Internet of Things (IoT), 5G edge computing, cloud network fusion and large-scale distributed deployment, the number of devices is growing exponentially.

[0003] Remote upgrade of these distributed devices for consistency, security and controllability has become a core difficulty in operation and maintenance management. The traditional large-scale upgrade method usually has the following problems, for example, the current common centralized or distributed upgrade scheme often pays attention to upgrade package distribution and version management, lacks scheduling strategies based on historical business load, and fails to fully utilize the business low peak window, thereby increasing the impact of upgrade on business. At the same time, different device hardware differences lead to different upgrade success rates, and a unified upgrade strategy cannot take into account the compatibility and stability of all devices. In the upgrade process, there is a lack of real-time stability evaluation and dynamic adjustment mechanism, which cannot timely perceive and handle abnormal upgrade, leading to large-scale rollback or non-recoverable problems. SUMMARY

[0004] In view of the defects of the prior art, the present application provides a method for controlling large-scale upgrade, which has the advantages of upgrading management for all devices and ensuring the stability of upgrade, and solves the above technical problems.

[0005] To achieve the above purpose, the present application provides the following technical scheme: a method for controlling large-scale upgrade, comprising the following steps: S1: obtaining an upgrade package and performing integrity verification based on the upgrade package, and then obtaining a hardware evaluation coefficient of each device; S2: after the integrity verification of the upgrade package is passed, executing an upgrade preparation instruction for all devices, specifically, constructing an upgrade stability coefficient of each device, constructing an overall upgrade stability coefficient by comprehensively considering the upgrade stability coefficient corresponding to the hardware evaluation coefficient of each device, and determining whether to execute S3 or issue a stability deviation instruction to execute a stability adjustment strategy; S3: reading a historical busy sequence list of the device, comprehensively calculating a business busy evaluation coefficient of all devices at each time period, and determining an upgrade start time; S4: issuing an upgrade instruction according to the upgrade start time, starting to execute an upgrade task for the corresponding device in batches, and constructing a device upgrade abnormality coefficient in the upgrade process. When the device upgrade abnormality coefficient exceeds a preset device upgrade abnormality threshold, an abnormality processing instruction is issued, and an abnormality processing strategy is executed; S5: After any batch of devices has been upgraded, perform a matching analysis on the currently upgraded devices and determine whether to trigger the rollback mechanism. If the rollback mechanism is triggered, continue analyzing the next batch until all devices have been upgraded.

[0006] As a preferred technical solution of the present invention, the step S1 of obtaining the upgrade package and performing integrity verification based on the upgrade package specifically includes the following steps: S1.a1: Get the version number of the current upgrade package and read the failure rate of the devices upgraded according to the current upgrade package in the cloud database. Specifically, it is the ratio of the total number of devices that have failed after the upgrade to the total number of upgraded devices stored in the cloud database. S1.a2: Read the probability of upgrade failure of the device being upgraded using the current upgrade package in the cloud database. Specifically, it is the ratio of the total number of devices that have encountered errors during the upgrade stored in the cloud database to the total number of upgrade devices stored in the cloud database. S1.a3: The current upgrade package update evaluation coefficient is constructed by weighting the failure rate of the devices being upgraded with the current upgrade package and the upgrade failure probability of the devices being upgraded with the current upgrade package. S1.a4: If the current upgrade package update evaluation coefficient is lower than the set update evaluation threshold, obtain the current upgrade package and perform integrity verification. If the current upgrade package update evaluation coefficient is not lower than the set update evaluation threshold, issue an early warning to the administrator, who will then decide whether to obtain the current upgrade package and perform integrity verification.

[0007] As a preferred embodiment of the present invention, the acquisition of the hardware evaluation coefficients for each device in step S1 is performed after step S1.a4 is completed, and the specific steps are as follows: S1.b1: Based on the first Obtain the hardware history data of the device. The hardware evaluation coefficient for each device is expressed as follows:

[0008] in, This represents the total number of hardware evaluation metrics. Indicates the first Standard values ​​for hardware evaluation metrics of a device Indicates the first The first device The average of the first hardware evaluation metric, based on the first Acquisition of historical hardware data for each device Represents absolute value. Indicates the first The first device Each hardware evaluation metric at the current moment corresponding specific value, indicates the hardware evaluation index of the device, indicates the hardware evaluation coefficient of the device, indicates summation; S1.b2: when the hardware evaluation coefficient of the device is lower than the hardware evaluation safety threshold value, a warning is given, and a worker is dispatched to overhaul the device.

[0009] As a preferred technical solution of the present application, the specific steps of evaluating the upgrade stability coefficient of each device in S2 are as follows: S2.a1: obtain the historical upgrade rollback coefficient of the device, specifically, the ratio of the historical upgrade rollback times of the device to the historical upgrade times of the device; S2.a2: obtain the backup verification evaluation coefficient of the device, specifically, assign a weight coefficient to the ratio of the number of verified data blocks to the total number of data blocks, and assign a weight coefficient to the ratio of the difference between the current time and the last backup time to the backup standard interval, and then add them to obtain the backup verification evaluation coefficient of the device; S2.a3: perform mean value operation on the inverse rate of the historical upgrade rollback coefficient of the device and the backup verification evaluation coefficient of the device to obtain the upgrade stability coefficient of the device.

[0010] As a preferred technical solution of the present application, the specific expression of constructing the overall upgrade stability coefficient by comprehensively considering the upgrade stability coefficient corresponding to the hardware evaluation coefficient of each device in S2 is as follows:

[0011] wherein, indicates the overall upgrade stability coefficient, indicates the total number of devices, indicates the hardware evaluation coefficient of the device, indicates summation.

[0012] ​As a preferred embodiment of the present invention, S2 further includes: when both the overall upgrade stability coefficient and the upgrade stability coefficient of each device exceed the corresponding set overall upgrade safety value and the set upgrade safety value, executing S3 if the overall upgrade stability coefficient... and the upgrade stability coefficient of each device If any value does not exceed the corresponding overall upgrade safety value or if none of the values ​​exceed the corresponding upgrade safety value, a stability deviation command is issued to execute the stability adjustment strategy.

[0013] As a preferred technical solution of the present invention, the stability adjustment strategy specifically includes the following steps: S2.b1: For the first The device was re-backed up, and the updated version was obtained. Upgrade stability coefficient of individual devices And will be the time The hardware evaluation indicators were eliminated, and the results were obtained after the staff performed the corresponding repairs. The first device Each hardware evaluation metric at time Corresponding specific value And based on S1.b1 and Reconstructing Moments The Hardware evaluation coefficient of individual devices ; S2.b2: Time-based The Hardware evaluation coefficient of individual devices And the updated Build Time Overall upgrade stability coefficient and based on Then reassess whether to issue a stabilization deviation command. If not, execute S3; if it is issued, the upgrade is paused.

[0014] As a preferred embodiment of the present invention, step S3 includes the following steps: S3.1: Read the historical busy sequence table of the devices and comprehensively calculate the business busy assessment coefficient of all devices for each time period. Specifically, it is the ratio of the total business volume of each time period to the sum of the total business volume of all time periods. S3.2: Get the upgrade duration of the current upgrade package, and then round up the ratio of the upgrade duration to the total duration of each time period to get the number of upgrade time periods; S3.3: Based on the number of upgrade periods, divide different periods into several time intervals in sequence, and sum the business busyness assessment coefficients of each period in each interval to obtain the business busyness assessment coefficient of the time interval; S3.4: After traversing all time interval, obtain the service busy evaluation coefficient of several time intervals, and select the starting point corresponding to the minimum time interval group as the upgrade starting time.

[0015] As a preferred technical solution of the present application, S4 specifically comprises the following steps: S4.1: Execute the upgrade instruction at the upgrade starting time to control all devices without executing tasks to upgrade in batches, and do not upgrade the devices executing tasks; S4.2: During the upgrade process, obtain, for the devices with update failure, execute re-upgrade, and record the upgrade times and the total upgrade time consumption , and construct the device upgrade abnormality coefficient , the specific expression is as follows:

[0016] Among them, indicates the maximum value of the upgrade times, indicates the maximum value of the total upgrade time consumption, indicates the upgrade abnormality, indicates the upgrade normality; S4.3: All device upgrade abnormality coefficients The devices are suspended from upgrading, an abnormality processing instruction is issued, and an abnormality processing strategy is executed; The abnormality processing strategy is specifically that dispatching workers to overhaul all suspended devices for upgrading, after the overhaul is completed, all overhauled devices and the devices completing the tasks in S4.1 are upgraded synchronously as the same batch, after the completion of the tasks of the devices in S4.1, no new tasks are dispatched, and the devices pre-upgraded in S4.1 will preferentially receive new tasks for processing.

[0017] As a preferred technical solution of the present application, S5 specifically comprises the following steps: S5.1: Construct the performance matching coefficient of the device after upgrade based on the deviation rate of the average response time before and after upgrade; S5.2: Sum the performance matching coefficients of the devices to obtain the comprehensive performance matching coefficient, when the comprehensive performance matching coefficient < 0, issue a performance warning, and remind the worker whether to trigger the rollback mechanism, when the comprehensive performance matching coefficient ≥ 0, complete the upgrade.

[0018] Compared with the prior art, the present application provides a method for controlling large-scale upgrade, which has the following beneficial effects: The application obtains the hardware evaluation coefficient of each device after verifying the upgrade package, constructs the upgrade stability coefficient of each device after the integrity verification of the upgrade package is passed, constructs the overall upgrade stability coefficient by comprehensively calculating the upgrade stability coefficient corresponding to the hardware evaluation coefficient of each device, and judges whether the stability adjustment strategy needs to be executed. Then the historical busy sequence list of the device is read, the business busy evaluation coefficient of each period of all devices is calculated, and the upgrade starting time is determined. After the upgrade of any batch of devices is completed, the matching analysis of the current upgrade completed device is performed, and it is judged whether the rollback mechanism is triggered. After the rollback mechanism is triggered, the next batch is continuously analyzed until all devices are upgraded, so that the one-time operation of large-scale upgrade management is realized, and the controllability of the upgrade process is ensured. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 The flowchart of the application is shown. DETAILED DESCRIPTION

[0020] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, not all. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the application.

[0021] Please refer to Figure 1 A method for controlling large-scale upgrade, comprising the following steps: S1: obtaining an upgrade package, and performing integrity verification based on the upgrade package, and then obtaining a hardware evaluation coefficient of each device; The step of obtaining the upgrade package and performing integrity verification based on the upgrade package in S1 specifically comprises the following steps: S1.a1: obtaining the version number of the current upgrade package, and reading the failure rate of the device performing upgrade according to the upgrade package in the cloud database The specific expression is as follows:

[0022] Among them, represents the total number of devices that appear fault error when running after upgrade stored in the cloud database, represents the total number of upgraded devices stored in the cloud database, and the failure condition of the remaining area during upgrade can be obtained from the cloud database to measure whether it will appear during the update process; S1.a2: reading the upgrade failure probability of the device performing upgrade in the cloud database The specific expression is as follows:

[0023] in, This indicates the total number of devices that encountered errors during the upgrade, stored in the cloud database. By collecting the percentage of the total number of devices that encountered errors during the upgrade, stored in the cloud database, we can reflect the stability of the current version of the upgrade package, and thus measure whether the current upgrade package will cause multiple device errors after the update. S1.a3: Comprehensively construct the current upgrade package update evaluation coefficients The specific expression is as follows:

[0024] in, , This represents two weight coefficients whose sum is 1; S1.a4: If the current upgrade package updates the evaluation coefficient Below the set update evaluation threshold At that time, obtain the current upgrade package and perform an integrity check. If the current upgrade package updates the evaluation coefficient... Not lower than the set update evaluation threshold When this happens, an alert is sent to the administrators, who then determine whether to acquire the current upgrade package and perform an integrity check. In this embodiment, the current upgrade package update evaluation coefficient is constructed. The parameters are shown in Table 1 below: Table 1 Constructs the evaluation coefficients for the current upgrade package update.

[0025] Calculation at this time It did not exceed the set update assessment threshold. Obtain the current upgrade package and perform integrity verification; The hardware evaluation coefficients for each device are obtained in S1 after S1.a4 is executed. The specific steps are as follows: S1.b1: Based on the first Obtain the hardware history data of the device. Hardware evaluation coefficient of individual devices The specific expression is as follows:

[0026] in, This represents the total number of hardware evaluation metrics. Indicates the first Standard values ​​for hardware evaluation metrics of a device Indicates the first The first device The average of the first hardware evaluation metric, based on the first Historical hardware data of each device Represents absolute value. Indicates the first The first device Each hardware evaluation metric at the current moment The corresponding specific value, Indicates the first The first device The median of the hardware evaluation indicators. For the specific hardware evaluation indicators of the i-th device, please refer to Table 2 below. (The specific parameters can be adjusted by those skilled in the art according to the type of equipment. For example, in the field of power equipment, the core indicators may be equipment temperature, voltage, current, etc., while in the field of UAV control, it may be the impact on the equipment.) Those skilled in the art can refer to Table 2 to make corresponding increases or decreases. Table 2 Hardware Evaluation Indicators

[0027] S1.b2: When the first Hardware evaluation coefficient of individual devices If the temperature falls below the hardware assessment safety threshold, an early warning will be issued, and staff will be dispatched to address the issue. Each piece of equipment was inspected and repaired. In this embodiment, the specific parameters are shown in Table 3 below: Table 3 constructs the first Hardware evaluation coefficient of individual devices

[0028] At this point, the calculation yields... The issue did not exceed the set hardware evaluation security threshold, so no warning was issued, and staff were dispatched to address it. Each piece of equipment was inspected and repaired. S2: After the integrity verification of the upgrade package is passed, the upgrade preparation instruction is executed on all devices. Specifically, the upgrade stability coefficient of each device is constructed, and the upgrade stability coefficient corresponding to the hardware evaluation coefficient of each device is combined to construct the overall upgrade stability coefficient. When the overall upgrade stability coefficient and the upgrade stability coefficient of each device both exceed the corresponding set overall upgrade safety value and the set upgrade safety value, S3 is executed. If either the overall upgrade stability coefficient or the upgrade stability coefficient of each device does not exceed the corresponding set overall upgrade safety value or neither exceeds the corresponding set upgrade safety value, a stability deviation instruction is issued and the stability adjustment strategy is executed. The specific steps for evaluating the upgrade stability coefficient of each device in S2 are as follows: S2.a1: Get the first Historical upgrade rollback coefficient for each device The specific expression is as follows:

[0029] in, Indicates the first Historical upgrade rollback coefficient for each device Indicates the first The number of historical upgrade rollbacks for each device. Indicates the first The number of historical upgrades for each device; S2.a2: Get the first Backup verification evaluation coefficient of each device The specific expression is as follows:

[0030] in, This indicates the number of data blocks that passed the verification. The data blocks verified here are all backup data blocks. Indicates the total number of data blocks. This indicates the current time, obtained directly from the server's system time. The timestamp from the backup data indicates the last backup time. Indicates the standard backup interval. and This represents two weighting coefficients that sum to 1. By obtaining the backup status of different devices before the update, it is possible that different devices may appear at different times due to differences in their operating cycles and internal parameters. Furthermore, the backup time of some devices may be modified or adjusted after maintenance, resulting in different backup times for different devices. S2.a3: Comprehensive construction of the first The stability coefficient for each device upgrade is expressed as follows:

[0031] in, Indicates the first The stability coefficient of each device upgrade It can be described as the first The reversal rate of the historical upgrade rollback coefficient for each device; In this embodiment, the first The specific parameters for the stability coefficient of each device upgrade are shown in Table 4 below: Table 4

[0032] The specific expression for constructing the overall upgrade stability coefficient by combining the upgrade stability coefficients corresponding to the hardware evaluation coefficients of each device in S2 is as follows:

[0033] in, This indicates the overall upgrade stability coefficient. This represents the total number of devices, in this embodiment. 0.87 exceeds the corresponding upgrade security value of 0.83; If the overall upgrade stability coefficient and the upgrade stability coefficient of each device If any value does not exceed the corresponding set overall upgrade safety value or if none of the values ​​exceed the corresponding set upgrade safety value, a stability deviation command is issued and a stability adjustment strategy is executed. The stability adjustment strategy specifically includes the following steps: S2.b1: For the first The device was re-backed up, and the updated version was obtained. Upgrade stability coefficient of individual devices And will be the time The hardware evaluation indicators were eliminated, and the results were obtained after the staff performed the corresponding repairs. The first device Each hardware evaluation metric at time Corresponding specific value and based on Reconstructing Moments The Hardware evaluation coefficient of individual devices Its specific construction method is similar to that above, and the specific expression is as follows:

[0034] in, Indicates the first The first device The median of the hardware evaluation metrics, Indicates the first The first device The average of each hardware evaluation metric, Indicates the first The first device Each hardware evaluation metric at time The corresponding specific value; S2.b2: Based on And the updated Build Time Overall upgrade stability coefficient The specific expression is as follows:

[0035] in, After that, based on And reassess whether to issue a stability deviation command. If not, execute S3. If it continues to issue, the upgrade is paused. In this case, after the stability adjustment strategy is triggered, the current value is retrieved again for replacement, and the corresponding device is backed up to ensure that a timely rollback can be made in case of subsequent failures. S3: Read the historical busy sequence table of devices, comprehensively calculate the business busy assessment coefficient of all devices for each time period, and determine the upgrade start time; S3 includes the following steps: S3.1: Read the historical busy sequence table of devices and calculate the business busyness evaluation coefficient of all devices for each time period. The specific expression is as follows:

[0036] in, This represents the business busyness assessment coefficient for the k-th time period. The total business volume for the k-th time period can be obtained from the historical average of the same period in the device's historical busy sequence table. This represents the summation of the total business volume across all time periods. This indicates the total number of time periods, where the time period is divided equally. S3.2: Get the upgrade duration of the current upgrade package and build upgrade time period numbers The specific expression is as follows:

[0037] in, This represents the total duration of the k-th time period. This indicates the upgrade duration of the current upgrade package. This indicates rounding up. The content here guarantees that the time period spanned by the upgrade corresponds to the k mentioned above, and that it is an integer. The minimum value is 1; S3.3: Based on the number of upgrade periods To obtain the business busyness assessment coefficient for the m-th time interval, the specific expression is as follows:

[0038] in, ,and This calculation process reflects the time period To the time period The total business activity in this range can be dynamically adjusted based on the upgrade duration of different upgrade packages, as described in this embodiment. This indicates that it needs to span three time periods, the specific time intervals are specified. This is used to measure the business activity in each time period. The update is performed when the interval is at its minimum, which minimizes the pressure on all devices. For example, time period 1, time period 2, and time period 3 form one interval, and time period 2, time period 3, and time period 4 form another interval. The final remainder is not considered. S3.4: After traversing all time intervals, obtain the business busyness assessment coefficients for several time intervals, and select the starting point corresponding to the set of time intervals with the smallest value as the upgrade start time. S4: Issue upgrade instructions based on the upgrade start time, start the corresponding devices to perform upgrade tasks in batches, and build a device upgrade anomaly coefficient during the upgrade process. When the device upgrade anomaly coefficient exceeds the preset device upgrade anomaly threshold, issue an anomaly handling instruction and execute the anomaly handling strategy. S4 specifically includes the following steps: S4.1: Execute the upgrade command at the start time of the upgrade, control all devices that have not performed tasks to perform the upgrade in batches, and do not upgrade devices that have performed tasks; S4.2: Obtain information during the upgrade process, perform a re-upgrade for devices that fail to update, and record the number of upgrades. Total upgrade time And construct the equipment upgrade anomaly coefficient, the specific expression of which is as follows:

[0039] in, This indicates the maximum number of upgrades. This represents the maximum total upgrade time. In real-world network environments, network conditions are complex and unpredictable. Poor network conditions can lead to slow downloads, disconnections, or upgrade failures. Therefore, it's necessary to limit the upgrade time for each device. This indicates an upgrade error. This indicates that the upgrade was successful. S4.3: Upgrade the anomaly coefficient of all devices. All equipment upgrades are suspended, an exception handling instruction is issued, and an exception handling strategy is executed. Specifically, the dispatch staff will inspect all the equipment whose upgrades have been suspended. After the inspection is completed, all the inspected equipment and the equipment that has completed the task in S4.1 will be upgraded synchronously in the same batch. After the equipment in S4.1 completes its task, no new task will be assigned. Equipment that has been upgraded in S4.1 will receive new tasks first and process them. S5: After any batch of devices has been upgraded, perform a matching analysis on the currently upgraded devices and determine whether to trigger the rollback mechanism. If the rollback mechanism is triggered, continue analyzing the next batch until all devices have been upgraded.

[0040] S5 specifically includes the following steps: S5.1: Obtain the upgraded version Performance matching coefficient of each device The specific expression is as follows:

[0041] in, Used to reflect performance metrics after the upgrade, when When this occurs, it indicates a performance degradation. When, it indicates a performance improvement. In this embodiment, and These represent the average response time before and after the upgrade, respectively. By measuring the degree of performance matching after the update, it reflects whether the updated device is more efficient. If the average response time shortens or remains unchanged in the later stages of the update, it indicates that the current device has better performance. Of course, those skilled in the art can also use different performance indicators for evaluation, as long as they reflect the relationship before and after the upgrade and meet the same proportional relationship, they can be replaced. For example, (memory utilization, transaction throughput, disk / network bandwidth utilization, IOPS, throughput). Those skilled in the art can make adaptive adjustments or replacements according to different devices. S5.2: Based on the first The overall performance matching coefficient is constructed from the performance matching coefficients of each device, and the specific expression is as follows:

[0042] in, This represents the overall performance matching coefficient. When the overall performance matching coefficient is... When the value is less than 0, a performance warning is issued, and staff are reminded whether to trigger the rollback mechanism. The above embodiments only provide one or more feasible solutions, and do not represent the optimal solution. The threshold size is set for ease of comparison. The size of the threshold depends on the amount of sample data and the number of bases set by those skilled in the art for each group of sample data. As long as it does not affect the ratio between the parameter and the quantized value, the weight can be determined by those skilled in the art based on each sample data and multiple rounds of experiments.

[0043] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method of controlling a mass upgrade, characterized by: The method comprises the following steps: S1: obtaining an upgrade package and performing integrity verification based on the upgrade package, and then obtaining a hardware evaluation coefficient of each device; S2: after the integrity verification of the upgrade package passes, executing upgrade preparation instructions on all devices, specifically, constructing an upgrade stability coefficient of each device, constructing an overall upgrade stability coefficient by synthesizing the upgrade stability coefficient corresponding to the hardware evaluation coefficient of each device, and determining whether to execute S3 or issue a stability deviation instruction to execute a stability adjustment strategy; S3: reading a historical busy sequence list of the device, synthesizing a service busy evaluation coefficient of all devices in each time period, and determining an upgrade start time; S4: issuing an upgrade instruction according to the upgrade start time, starting to execute an upgrade task on the corresponding device in batches, and constructing a device upgrade abnormality coefficient in the upgrade process; when the device upgrade abnormality coefficient exceeds a preset device upgrade abnormality threshold, issuing an abnormality processing instruction and executing an abnormality processing strategy; S5: when the upgrade of any batch of devices is completed, performing matching analysis on the current upgraded devices, and determining whether to trigger a rollback mechanism; after the rollback mechanism is triggered, the next batch is continuously analyzed until the upgrade of all devices is completed.

2. The method of claim 1, wherein: The step of obtaining the upgrade package and performing integrity verification based on the upgrade package in S1 specifically comprises the following steps: S1.a1: obtaining a version number of the current upgrade package, and reading a failure rate of the device upgraded based on the current upgrade package from the cloud database, specifically, a ratio of a total number of devices that have reported errors when running after being upgraded and stored in the cloud database to a total number of upgraded devices stored in the cloud database; S1.a2: reading an upgrade failure probability of the device upgraded based on the current upgrade package from the cloud database, specifically, a ratio of a total number of devices that have reported errors during the upgrade and stored in the cloud database to a total number of upgraded devices stored in the cloud database; S1.a3: synthesizing the failure rate of the device upgraded based on the current upgrade package and the upgrade failure probability of the device upgraded based on the current upgrade package, and constructing a current upgrade package update evaluation coefficient by weighting; S1.a4: if the current upgrade package update evaluation coefficient is lower than a set update evaluation threshold, obtaining the current upgrade package and performing integrity verification; if the current upgrade package update evaluation coefficient is not lower than the set update evaluation threshold, issuing a warning prompt to the manager to determine whether to obtain the current upgrade package and perform integrity verification.

3. A method of controlling mass upgrades according to claim 2, characterized in that: The step of obtaining the hardware evaluation coefficient of each device in S1 is executed after S1.a4 is executed, and the specific steps are as follows: S1.b1: Based on the first Obtain the hardware history data of the device. The hardware evaluation coefficient for each device is expressed as follows: wherein, represents the total number of hardware evaluation indicators, represents the standard value of the hardware evaluation indicator of the device, represents the mean value of the hardware evaluation indicator of the device, based on the hardware historical data of the device, represents the absolute value, represents the specific value of the hardware evaluation indicator of the device at the current time, corresponding, represents the median of the hardware evaluation indicator of the device, represents the hardware evaluation coefficient of the device, represents summation; S1.b2: when the hardware evaluation coefficient of the first device is lower than the hardware evaluation safety threshold, a warning is given, and a worker is dispatched to overhaul the first device. device.

4. The method of controlling large-scale upgrades of claim 3, wherein: The specific steps of evaluating the upgrade stability coefficient of each device in S2 are as follows: S2.a1: obtaining the first device historical upgrade rollback coefficient, specifically, the ratio of the first device historical upgrade rollback times to the first device historical upgrade times; S2.a2: Get the first The backup verification evaluation coefficient for each device is specifically a weighted coefficient assigned to the ratio of the number of verified data blocks to the total number of data blocks. The ratio of the difference between the current time and the last backup time to the standard backup interval is then used to assign a weighting coefficient. After adding them together, we get the first one. Backup verification evaluation coefficient for each device; S2.a3: the reverse rate of the device history upgrade rollback coefficient of the first device is subjected to mean operation with the backup check evaluation coefficient of the first device to obtain the upgrade stability coefficient of the first device.

5. A method of controlling mass upgrades according to claim 4, characterized in that: The specific expression of constructing the overall upgrade stability coefficient by synthesizing the upgrade stability coefficient corresponding to the hardware evaluation coefficient of each device in S2 is as follows: wherein, represents the overall upgrade stability coefficient, represents the total number of devices, represents the hardware evaluation coefficient of the device, represents the summation.

6. A method of controlling mass upgrades according to claim 5, wherein: The S2 further comprises: when the overall upgrade stability coefficient and the upgrade stability coefficient of each device all exceed the corresponding set overall upgrade safety value and set upgrade safety value, performing S3 if the overall upgrade stability coefficient and the upgrade stability coefficient of each device Any one does not exceed the corresponding set overall upgrade safety value or does not exceed the corresponding set upgrade safety value, then a stable deviation instruction is issued, and a stability adjustment strategy is performed.

7. A method of controlling mass upgrades according to claim 6, characterized in that: The stability adjustment strategy specifically comprises the following steps: S2.b1: For the first The device was re-backed up, and the updated version was obtained. Upgrade stability coefficient of individual devices And will be the time The hardware evaluation indicators were eliminated, and the results were obtained after the staff performed the corresponding repairs. The first device Each hardware evaluation metric at time Corresponding specific value And based on S1.b1 and Reconstructing Moments The Hardware evaluation coefficient of individual devices ; S2.b2: based on the hardware evaluation coefficient of the first device at the time S2.b3: based on the overall upgrade stability coefficient of the first device at the time of construction S2.b4: based on the re-updated hardware evaluation coefficient of the first device at the time of construction S2.b5: based on the re-updated overall upgrade stability coefficient of the first device at the time of construction S2.b6: based on the re-updated hardware evaluation coefficient of the second device at the time of construction S2.b7: based on the re-updated overall upgrade stability coefficient of the second device at the time of construction S2.b8: based on the re-updated hardware evaluation coefficient of the third device at the time of construction S2.b9: based on the re-updated overall upgrade stability coefficient of the third device at the time of construction S 8. The method of controlling large-scale upgrades of claim 1, wherein: S3 comprises the following steps: S3.1: reading a historical busy sequence list of the device, synthesizing a service busy evaluation coefficient of all devices in each time period, specifically, a ratio of a total amount of services in each time period to a total amount of services in all time periods; S3.2: Obtain the upgrade duration of the current upgrade package, and obtain the upgrade time period number by dividing the total duration of each time period and rounding up; S3.3: Based on the upgrade time period number, sequentially divide the different time periods into a plurality of time period intervals, and sum the service busy evaluation coefficients of each time period in each interval to obtain the service busy evaluation coefficients of the time period interval; S3.4: After traversing all the time period intervals, obtain the service busy evaluation coefficients of the plurality of time period intervals, and select the starting point corresponding to the smallest group of time period intervals as the upgrade starting time.

9. A method of controlling mass upgrades according to claim 8, characterized in that: The S4 specifically comprises the following steps: S4.1: Execute the upgrade instruction at the upgrade starting time to control all devices that have not executed the task to perform batch upgrade, and do not upgrade the devices that have executed the task; S4.2: In the upgrading process, for the device that fails to update, perform re-upgrade and record the number of upgrades and the total time of upgrading and construct the device upgrading abnormality coefficient The specific expression is as follows: wherein, represents the maximum value of the number of upgrades, represents the maximum value of the total upgrade time, represents an upgrade anomaly, represents a normal upgrade; S4.3: Upgrade all devices with abnormal coefficients All devices suspend upgrading, issue abnormal processing instructions, and execute abnormal processing strategies; The abnormal processing strategy is specifically to dispatch workers to repair all the devices that have been suspended from upgrading, after the repair is completed, all the repaired devices and the devices that have completed the task in S4.1 are upgraded synchronously as the same batch, after the task of the devices in S4.1 is completed, no new task is dispatched, and the devices that have been pre-upgraded in S4.1 will receive the new task for processing in priority.

10. The method of controlling large-scale upgrades of claim 9, wherein: The S5 specifically comprises the following steps: S5.1: constructing the performance matching coefficient of the device after the upgrade based on the deviation rate of the average response time before the upgrade and after the upgrade; S5.1: constructing the performance matching coefficient of the device after the upgrade based on the deviation rate of the average response time before the upgrade and after the upgrade; S5.2: Sum the performance matching coefficients of the first devices to obtain a comprehensive performance matching coefficient, when the comprehensive performance matching coefficient < 0, send a performance warning, and remind the staff whether to trigger the fallback mechanism, when the comprehensive performance matching coefficient ≥ 0, complete the upgrade.

Images