A resource management method for a virtual machine system

Abstract

The present invention discloses a resource management method for a virtual machine system. The method first obtains the usage habit data of all users of all virtual machines, then selects a virtual machine currently being used by a user, performs inertia analysis on the usage habit data, mainly determines the deviation value of the usage upper value Zi of each day in the accumulation stage according to the usage upper value, determines a number of removal upper values ​​and a number of range upper values ​​according to the relationship between the deviation value and X1, determines the removal ratio according to the number of removal upper values, then automatically obtains corresponding one of combined information, semi-combined information and mismatched information, and then automatically performs storage allocation; after the allocation is completed, performs real-time analysis on the use of the virtual machine by the user, combines the daily average load of the CPU and the storage change ratio to determine whether the user is in an idle state when using the virtual machine, and automatically reminds the management personnel to handle when the judgment condition is met, thereby avoiding complicated alarm signals and accurately identifying.

Description

Technical Field

[0001] The present invention relates to the technical field of asset management, and in particular to a resource management method for a virtual machine system. Background Art

[0002] In the existing technology, virtual machines can be monitored and warned through vRealize Operations Manager (vROps), a centralized management and monitoring product of VMware's virtualization platform, which enables operation and maintenance personnel to discover and solve potential risk problems in advance. For example, it can issue warnings for busy virtual machines, reminding operation and maintenance personnel to expand resources, and it will also issue alarms for idle virtual machines, reminding operation and maintenance personnel to recycle resources. These are all very necessary and practical functions.

[0003] However, in actual use, it is found that there are many problems, which make it impossible to effectively manage virtualized resources, such as:

[0004] First, when allocating resources, especially for users, the user's pre-allocation method is generally as follows: For example, when the user actually purchases 20 TB, but when it is actually allocated, part of it will be allocated first, such as 1 TB. For some users, if the business bursts and requires storage content beyond the scope, the user's behavior cannot be optimized. Therefore, how to reasonably allocate appropriate resources to users and optimize user experience is a difficult problem;

[0005] Secondly, there are too many alarm messages and they cannot be processed automatically; the false alarm rate of idle virtual machines is too high and they cannot be restored after being recycled.

[0006] The reason for the above problems is that the centralized management and monitoring products in the existing technology only use performance data as indicators for evaluation. The evaluation method is single and generates a large amount of alarm information for operation and maintenance personnel to handle, and the processing efficiency is low. Summary of the invention

[0007] The purpose of the present invention is to provide a resource management method for a virtual machine system;

[0008] The purpose of the present invention can be achieved through the following technical solutions:

[0009] A resource management method for a virtual machine system, the method specifically comprising the following steps:

[0010] Step 1: Obtain the usage habit data of all users of all virtual machines. The usage habit data is the highest storage space usage value of the user every day during the accumulation stage.

[0011] Choose a virtual machine that a user is currently using, and perform inertial analysis on its usage habit data. Mainly, for the daily usage upper value Zi in the accumulation stage, determine its deviation value according to the usage upper value, determine several removal upper values ​​and several range upper values ​​according to the relationship between the deviation value and X1, determine the removal ratio according to the number of removal upper values, and generate any one of the combined signal, semi-combined signal and mismatched signal according to the removal ratio, and then allocate regular storage value and extended upper value according to different signals to obtain the corresponding one of the combined information, semi-combined information and mismatched information;

[0012] Step 2: Automatically perform storage allocation according to any of the conventional information, semi-closed information, and misclosed information obtained. The specific storage allocation method is as follows:

[0013] S01: When the obtained information is regular information, the regular stored value and the extended value in the regular information are automatically obtained;

[0014] Allocate the user's next stage of allocated storage space as regular storage value, and conduct real-time analysis of the user's usage in the next stage. When the real-time usage space occupies 95% of the regular storage value, quickly add additional space to bring the total storage capacity of the space within the extended upper value range;

[0015] S02: When the semi-closed information is obtained, the normal storage value and the extended upper value in the semi-closed information are automatically obtained; the allocated storage space of the user in the next stage is allocated as the normal storage value, and the user's usage in the next stage is analyzed in real time. When the real-time usage space occupies 80% of the normal storage value, additional space is quickly added to bring the total storage capacity of the space to within the extended upper value range;

[0016] S03: When mismatch information is obtained, the normal storage value and the extended upper value in the mismatch information are automatically obtained; the allocated storage space of the user in the next stage is allocated as the normal storage value, and the user's usage in the next stage is analyzed in real time. When the real-time usage space occupies 60% of the normal storage value, additional space is quickly added to bring the total storage capacity of the space to within the extended upper value range;

[0017] S04: Complete the allocation and quick call of storage space.

[0018] Beneficial effects of the present invention:

[0019] The present invention first obtains the usage habit data of all users of all virtual machines, then selects a virtual machine currently being used by any user, performs inertial analysis on the usage habit data, mainly determines the deviation value of the usage upper value Zi of each day in the accumulation stage according to the usage upper value, determines a number of removal upper values ​​and a number of range upper values ​​according to the relationship between the deviation value and X1, determines the removal ratio according to the number of removal upper values, and generates any one of the combined signal, semi-combined signal and mismatched signal according to the removal ratio, and then allocates conventional storage value and extended upper value according to different signals to obtain corresponding one of the combined information, semi-combined information and mismatched information; and then automatically performs storage allocation according to any one of the obtained conventional information, semi-combined information and mismatched information;

[0020] After the allocation is completed, the user's use of the virtual machine is analyzed in real time, combined with the daily average CPU load and storage change ratio to determine whether the user is idle when using the virtual machine. When the judgment conditions are met, the management personnel are automatically reminded to handle it, avoiding complicated alarm signals and accurately identifying them. DETAILED DESCRIPTION

[0021] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] The present invention is a resource management method for a virtual machine system, which specifically comprises the following steps:

[0023] Step 1: Obtain the usage habit data of all users of all virtual machines. The usage habit data is the highest storage space usage value of the user every day in the accumulation stage. The storage space usage value refers to the maximum storage space used by the user on that day. The accumulation stage specifically refers to the stage of 180 days from the current time of inertia analysis.

[0024] Select any virtual machine that a user is currently using and perform inertia analysis on its usage habit data. The specific method of inertia analysis is as follows:

[0025] S1: Obtain the highest storage space usage value of each day in the cumulative stage of the usage habit data, and mark it as the usage value Zi, where i=1, ..., 180;

[0026] S2: Then automatically calculate the mean of Zi, mark it as P, and use the formula to calculate the deviation value W of the upper value. The specific calculation formula is:

[0027]

[0028] In the formula, |*| means taking the absolute value of the value in brackets;

[0029] S3: Then the W value is determined. When W exceeds X1, X1 is a preset value, and data is deleted. The specific method of data deletion is as follows:

[0030] Sort Zi in descending order according to |Zi-P|, select the corresponding Zi in turn, remove them temporarily, and recalculate the W value of the remaining Zi after removal. If the W value still exceeds X1 at this time, select the next Zi value in descending order according to |Zi-P|, remove it and recalculate the W value until the W value does not exceed X1;

[0031] S4: Mark the removed Zi values ​​as removed upper values, and the remaining ones as range upper values;

[0032] S5: Determine according to the number of removed upper values, obtain the number of removed upper values, divide it by 180, and obtain the removal ratio;

[0033] When the removal ratio is lower than X2, a combined signal is generated;

[0034] When the removal ratio is between X2 and 4*X2, a semi-closed signal is generated;

[0035] Otherwise, a mismatch signal is generated;

[0036] Here X2 is a preset value, which is generally 10%;

[0037] S6: Allocate the conventional storage value and the extended upper value according to the generated closing signal, half-closing signal and mismatching signal. The specific allocation method is as follows:

[0038] When the combined signal is generated, the maximum value of the range upper value is marked as the regular storage value, and the maximum value of the removed upper value is marked as the extended upper value; the information formed by the fusion of the regular storage value and the extended upper value at this time is marked as regular information;

[0039] When a semi-closed signal is generated, the intermediate data in the removed upper value is automatically divided into the range upper value, and the number of intermediate data is half of the number of removed upper values. Specifically, the removed upper values ​​are sorted in order from large to small, and the data content with the middle sorting of 25% to 75% is taken; then the range upper value and the removed upper value are updated, and the maximum value of the range upper value at this time is marked as the regular storage value, and the maximum value of the removed upper value is marked as the extended upper value; the information formed by the fusion of the regular storage value and the extended upper value at this time is marked as semi-closed information;

[0040] When a mismatch signal is generated, all the removed upper values ​​and range upper values ​​are automatically obtained and merged. After fusion, the mean is automatically calculated, and the median between the mean and the maximum value of all fused data is marked as the regular storage value. The value obtained by multiplying the maximum value of all fused data by X3 is marked as the extended upper value, where X3 is a preset value, usually 1.25; the information formed by the fusion of the regular storage value and the extended upper value at this time is marked as mismatch information;

[0041] S7: obtaining any one of conventional information, semi-matching information, and mismatching information according to the above analysis content;

[0042] Step 2: Automatically perform storage allocation according to any of the conventional information, semi-closed information, and misclosed information obtained. The specific storage allocation method is as follows:

[0043] S01: When the obtained information is regular information, the regular stored value and the extended value in the regular information are automatically obtained;

[0044] Allocate the user's next stage of allocated storage space as regular storage value, and conduct real-time analysis of the user's usage in the next stage. When the real-time usage space occupies 95% of the regular storage value, quickly add additional space to bring the total storage capacity of the space within the extended upper value range;

[0045] S02: When the semi-closed information is obtained, the normal storage value and the extended upper value in the semi-closed information are automatically obtained; the allocated storage space of the user in the next stage is allocated as the normal storage value, and the user's usage in the next stage is analyzed in real time. When the real-time usage space occupies 80% of the normal storage value, additional space is quickly added to bring the total storage capacity of the space to within the extended upper value range;

[0046] S03: When mismatch information is obtained, the normal storage value and the extended upper value in the mismatch information are automatically obtained; the allocated storage space of the user in the next stage is allocated as the normal storage value, and the user's usage in the next stage is analyzed in real time. When the real-time usage space occupies 60% of the normal storage value, additional space is quickly added to bring the total storage capacity of the space to within the extended upper value range;

[0047] S04: complete the allocation and quick call of storage space;

[0048] Step 3: Monitor the user's usage. The specific methods of supervision are:

[0049] SS1: Collect the operation data of the virtual machines used by users. The specific method of operation data collection is as follows:

[0050] Get the average CPU load and storage change ratio of the virtual machine every day; the specific method of obtaining the storage change ratio is:

[0051] The storage space usage of the current day is obtained and marked as the real-time usage; then the storage space usage of the previous day is obtained and marked as the past usage; the real-time usage is subtracted from the past usage and divided by the real-time usage, and marked as the storage change ratio;

[0052] SS2: First, remove the dimensions of the average load and storage change ratio and mark them as Pf and Cb respectively. Use the formula to calculate the daily usage value D. The specific calculation formula is:

[0053] D = 0.54*Pf + 0.46*Cb;

[0054] In the formula, * represents a multiplication sign;

[0055] SS3: Then, the daily usage value of the user when using the virtual machine is continuously obtained, and a daily usage value group is continuously obtained, which is marked as Dj, j=1...n; Dn represents the daily usage value at the latest time, and n is a positive integer;

[0056] SS4: Use the formula to calculate the change value Bdj. The specific calculation formula is:

[0057] B j =D j+1 -D j , j=1, ..., n-1;

[0058] SS5: Get the value in Bdj that is less than X4, divide it by n-1, and mark the resulting value as the base value; X4 is a preset value, usually 3;

[0059] SS6: When the bottom value exceeds X5, a low stop signal is generated. X5 is a preset value, which is generally 0.75.

[0060] Step 4: When a low-stop signal is generated, the administrator is automatically prompted that the current user's virtual machine usage is idle and needs to be processed as soon as possible;

[0061] Step 5: Complete the resource management method.

[0062] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A resource management method for a virtual machine system, characterized in that: The method specifically comprises the following steps: Step 1: Obtain the usage habit data of all users of all virtual machines. The usage habit data is the highest storage space usage value of the user every day during the accumulation stage. Choose a virtual machine that a user is currently using, and perform inertial analysis on its usage habit data. Mainly, for the daily usage upper value Zi in the accumulation stage, determine its deviation value according to the usage upper value, determine several removal upper values ​​and several range upper values ​​according to the relationship between the deviation value and X1, determine the removal ratio according to the number of removal upper values, and generate any one of the combined signal, semi-combined signal and mismatched signal according to the removal ratio, and then allocate regular storage value and extended upper value according to different signals to obtain the corresponding one of the combined information, semi-combined information and mismatched information; Step 2: Automatically perform storage allocation according to any of the conventional information, semi-closed information, and misclosed information obtained. The specific storage allocation method is as follows: S01: When the obtained information is regular information, the regular stored value and the extended value in the regular information are automatically obtained; Allocate the user's next stage of allocated storage space as regular storage value, and conduct real-time analysis of the user's usage in the next stage. When the real-time usage space occupies 95% of the regular storage value, quickly add additional space to bring the total storage capacity of the space within the extended upper value range; S02: when the obtained information is semi-closed, the conventional stored value and the extended upper value in the semi-closed information are automatically obtained; Allocate the user's next stage of allocated storage space as regular storage value, and conduct real-time analysis of the user's usage in the next stage. When the real-time usage space occupies 80% of the regular storage value, quickly add additional space to bring the total storage capacity of the space within the extended upper value range; S03: when mismatch information is obtained, automatically obtain the normal stored value and the extended upper value in the mismatch information; Allocate the user's next stage of allocated storage space as regular storage value, and conduct real-time analysis of the user's usage in the next stage. When the real-time usage space occupies 60% of the regular storage value, quickly add additional space to bring the total storage capacity of the space within the extended upper value range; S04: Complete the allocation and quick call of storage space.

2. A method for managing resources of a virtual machine system according to claim 1, characterized in that: The storage space usage value refers to the maximum storage space used by the user on that day; the cumulative stage specifically refers to the stage of 180 days from the current time of inertial analysis.

3. A method for managing resources of a virtual machine system according to claim 1, characterized in that: The specific method of inertia analysis in step 1 is: S1: Obtain the highest storage space usage value of each day in the cumulative stage of the usage habit data, and mark it as the usage value Zi, where i=1, ..., 180; S2: Then automatically calculate the mean of Zi, mark it as P, and use the formula to calculate the deviation value W of the upper value. The specific calculation formula is: In the formula, |*| means taking the absolute value of the value in brackets; S3: Then the W value is determined. When W exceeds X1, X1 is a preset value, and data is deleted. The specific method of data deletion is as follows: Sort Zi in descending order according to |Zi-P|, select the corresponding Zi in turn, remove them temporarily, and recalculate the W value of the remaining Zi after removal. If the W value still exceeds X1 at this time, select the next Zi value in descending order according to |Zi-P|, remove it and recalculate the W value until the W value does not exceed X1; S4: Mark the removed Zi values ​​as removed upper values, and the remaining ones as range upper values; S5: Determine according to the number of removed upper values, obtain the number of removed upper values, divide it by 180, and obtain the removal ratio; When the removal ratio is lower than X2, a combined signal is generated; When the removal ratio is between X2 and 4*X2, a semi-closed signal is generated; Otherwise, a mismatch signal is generated; S6: According to the generated combined signal, semi-combined signal and mismatched signal, a conventional stored value and an extended upper value are allocated, and any one of the conventional information, semi-combined information and mismatched information is obtained according to the above analysis content.

4. A method for managing resources of a virtual machine system according to claim 3, characterized in that: The specific method of allocating the conventional stored value and the extended upvalue in step S6 is: When the combined signal is generated, the maximum value of the range upper value is marked as the regular storage value, and the maximum value of the removed upper value is marked as the extended upper value; the information formed by the fusion of the regular storage value and the extended upper value at this time is marked as regular information; When a semi-closed signal is generated, the intermediate data in the removed upper value is automatically divided into the range upper value, and the number of intermediate data is half of the number of removed upper values. Specifically, the removed upper values ​​are sorted in order from large to small, and the data content with the middle sorting of 25% to 75% is taken; then the range upper value and the removed upper value are updated, and the maximum value of the range upper value at this time is marked as the regular storage value, and the maximum value of the removed upper value is marked as the extended upper value; the information formed by the fusion of the regular storage value and the extended upper value at this time is marked as semi-closed information; When a mismatch signal is generated, all the removed upper values ​​and range upper values ​​are automatically obtained and merged. After the fusion, the mean is automatically calculated, and the median between the mean and the maximum value of all the data after the fusion is marked as the regular storage value. The value obtained by multiplying the maximum value of all the data after the fusion by X3 is marked as the extended upper value, where X3 is a preset value, usually 1.25; the information formed by the fusion of the regular storage value and the extended upper value at this time is marked as mismatch information.

5. The resource management method of a virtual machine system according to claim 3, characterized in that: X2 in step S5 is a preset value.

6. A method for managing resources of a virtual machine system according to claim 3 or 5, characterized in that: The specific value of X2 is 10%.

7. A method for managing resources of a virtual machine system according to claim 1, characterized in that: After completing step 2, you need to perform the following steps, specifically: Step 3: Monitor the user's usage. The specific methods of supervision are: SS1: Collect the operation data of the virtual machines used by users. The specific method of operation data collection is as follows: Get the average CPU load and storage change ratio of the virtual machine every day; SS2: First, remove the dimensions of the average load and storage change ratio and mark them as Pf and Cb respectively. Use the formula to calculate the daily usage value D. The specific calculation formula is: D = 0.54*Pf + 0.46*Cb; In the formula, * represents a multiplication sign; SS3: Then, the daily usage value of the user when using the virtual machine is continuously obtained, and a daily usage value group is continuously obtained, which is marked as Dj, j=1...n; Dn represents the daily usage value at the latest time, and n is a positive integer; SS4: Use the formula to calculate the change value Bdj. The specific calculation formula is: Bd j =D j+1 -D j ,j=1、...、n-1; SS5: Get the value in Bdj that is less than X4, divide it by n-1, and mark the resulting value as the base value; SS6: When the bottom value exceeds X5, a low stop signal is generated; Step 4: When a low-stop signal is generated, the administrator is automatically prompted that the current user's virtual machine usage is idle and please handle it as soon as possible.

8. A method for managing resources of a virtual machine system according to claim 7, characterized in that: The specific method of obtaining the storage change ratio in step SS1 is: The storage space usage of the day is obtained and marked as the real-time usage; then the storage space usage of the previous day is obtained and marked as the past usage; the real-time usage minus the past usage is divided by the real-time usage and marked as the storage change ratio.

9. A method for managing resources of a virtual machine system according to claim 7, characterized in that: X4 is the preset value; X5 is the preset value.

Images