JP5446667B2 - Virtual computer management system, virtual computer management method and program - Google Patents

Description

  The present invention relates to a virtual computer management system, a virtual computer management method, and a program for monitoring an operating state of a virtual computer to which computer resources are dynamically allocated.

  By using virtual machine (VM) technology, it is possible to operate a computer system beyond hardware limitations, such as server creation and replication, movement of operating locations, and dynamic processing performance changes. Can be done easily. For this reason, the virtualization technology is expected to be able to realize flexible service provision following changes in the system environment.

  In addition, since the virtual machine behaves logically in the same way as the physical machine, the server (guest OS) incorporated in the virtual machine measures the load such as the CPU usage rate and the number of job processes in the virtual machine, It has a function of displaying the load history in a graph.

  However, usually, when the setting change of the resource allocation amount for the virtual machine is dynamically performed, the load amount such as the CPU usage rate and job processing rate of the virtual machine is increased or discounted with the change of the resource allocation amount. Therefore, there is a problem that the continuity of the graph is lost when the load history is displayed in a graph.

  As a result, it has been difficult to determine whether the load fluctuation factor of the virtual machine is outside the virtual machine, such as setting change of the resource allocation amount, or inside the virtual machine, such as a problem of the execution application. This makes it impossible to set a threshold value for determining an abnormality of a virtual machine, making it difficult to determine a failure occurring in the virtual machine, causing a false detection of a failure or a detection failure, There has been a problem that the efficiency of the operation management work is greatly reduced. As a factor outside the virtual machine, if the virtual machine is moved to a physical machine with different performance, the performance of the virtual machine is changed, the resource allocation to the virtual machine is changed, the virtual machine on the physical machine There are cases where the number has changed.

  Therefore, a related technique for this is disclosed in Patent Document 1. In Patent Document 1, the actual elapsed time counter is used on the guest OS side to measure the CPU usage time that is not assigned to the guest OS, and the measured CPU time is regarded as idle and the physical CPU usage of the guest OS in the entire physical CPU. Techniques for measuring rates are disclosed. The physical CPU usage rate of the guest OS is information that does not increase or decrease even if the CPU allocation rate for the guest OS is changed.

JP 2008-225655 A

  However, the related technique disclosed in Patent Document 1 is applied to the CPU usage rate among the loads measured by the guest OS, and is not applicable to values represented by absolute numbers such as the number of job processes. There was a problem that I could not. In addition, in order to calculate the physical CPU usage rate, it is necessary to measure the CPU time not allocated to the guest OS, and the calculation processing of the physical CPU usage rate is performed separately from the processing for calculating the normal virtual CPU usage rate. Since this has to be done, there has been a problem that the processing load increases.

  Therefore, the present invention solves the problems associated with the related art, and as information for monitoring a virtual machine, information that can represent the load of the virtual machine without being affected by the resource allocation amount to the virtual machine It is an object of the present invention to provide a virtual machine management system, a virtual machine management method, and a program for obtaining the above.

In order to achieve the above object, a virtual machine management system according to the present invention is a virtual machine management system that monitors an operating state of a guest machine that is a virtual machine to which computer resources are dynamically allocated by a host computer. Performance information acquisition means for acquiring performance information indicating the load of the computer, resource allocation information acquisition means for acquiring resource allocation information indicating the amount of computer resources pre-allocated to the guest computer from the host computer side, and this resource Performance information correcting means for correcting the performance information acquired by the performance information acquiring means based on the resource allocation information acquired by the allocation information acquiring means and outputting the corrected information as monitoring data; and the performance information correcting means Graph data which is image data of a monitoring graph representing the history of the monitoring data by inputting the monitoring data from Includes a graph generating means for generating for the said graph generating means, the type of resource allocation information from resource assignment information obtaining unit, the surveillance based on the resource allocation amount to the guest computer indicating the resource allocation information The graph has a function of compressing or expanding the scale interval on the time axis in the graph and converting the image data of the graph into the graph data .

In the virtual machine management method of the present invention, the performance information acquisition unit acquires the performance information indicating the load of the guest computer that is a virtual machine to which the computer resources are dynamically allocated by the host computer. In addition, resource allocation information acquisition means indicating the amount of computer resources allocated in advance to the guest computer is acquired from the host computer side by the resource allocation information acquisition means, and the adjustment corresponding to the resource allocation amount indicated by the acquired resource allocation information the performance information by correcting performance information correction hands stage by multiplying the load value indicated by the acquired performance information magnification, and outputs the information after the correction as monitoring data for monitoring the guest computer, said Based on the resource allocation amount indicated by the resource allocation information, the graph generation means indicates the size of the time axis scale interval in the monitoring graph representing the history of the monitoring data. Contraction or expansion, and generating said graph generating means graph data is image data of the graph.

Furthermore, the virtual machine management program of the present invention includes a performance information acquisition function for acquiring performance information indicating a load of a guest computer that is a virtual computer to which computer resources are dynamically allocated by a host computer, and a pre-stored in the guest computer. A resource allocation information acquisition function for acquiring resource allocation information indicating the amount of allocated computer resources, and a load value indicated by the acquired performance information indicating an adjustment factor corresponding to the resource allocation amount indicated by the acquired resource allocation information Based on the resource allocation amount indicated by the resource allocation information and the performance information correction function for correcting the performance information and outputting the corrected information as monitoring data for monitoring the guest computer In the monitoring graph representing the history of the monitoring data, the size of the scale interval on the time axis is compressed or expanded, and the image data of this graph Characterized in that to achieve the graph generation function of generating graph data, to the computer.

  Since the present invention is configured as described above, this corrects the load value of the virtual machine in accordance with the resource allocation amount for the virtual machine, and uses the corrected information as monitoring data. A value obtained by canceling the increase / decrease related to the resource allocation amount to the virtual machine from the load amount can be obtained as monitoring data. By using this monitoring data, it is possible to easily execute a fault determination for the virtual machine. it can.

It is a functional block diagram which shows the structure of the virtual machine management system of one Embodiment concerning this invention. (A) is a graph showing the history of the CPU usage rate of the guest computer corrected by the performance information correcting unit in the embodiment disclosed in FIG. (B) is a graph showing the history of the CPU usage rate of the guest computer acquired by the performance information acquisition means in the embodiment disclosed in FIG. (C) is a graph showing the history of the CPU allocation rate to the guest computer acquired by the resource allocation information acquisition unit in the embodiment disclosed in FIG. It is a figure which shows the correspondence of adjustment magnification | multiplying_factor and resource allocation in embodiment disclosed in FIG. 1 about the case where resource allocation information is a CPU allocation rate. (A) is a graph showing the history of the number of processed jobs of the guest computer corrected by the performance information correcting unit in the embodiment disclosed in FIG. (B) is a graph showing the history of the number of job processes of the guest computer acquired by the performance information acquisition unit in the embodiment disclosed in FIG. (C) is a graph showing the history of the CPU allocation rate to the guest computer acquired by the resource allocation information acquisition unit in the embodiment disclosed in FIG. 2 is a flowchart showing the operation of the virtual machine management system of the embodiment disclosed in FIG. 1.

  Hereinafter, an embodiment according to the present invention will be described with reference to the drawings.

  FIG. 1 is a functional block diagram showing the configuration of the virtual machine management system 1 of this embodiment. As shown in FIG. 1, the virtual machine management system 1 according to the present embodiment is a system that monitors the operating state of a guest computer 22 that is a virtual machine provided on a host computer 2.

  The host computer 2 has a function of dynamically allocating computer resources under the control of the host computer 2 to the guest computer 22. Here, the computer resources are hardware resources constituting the computer, such as a CPU (Central Processing Unit), a main memory, and an I / O (Input / Output).

  As shown in FIG. 1, the host computer 2 has a resource allocation collection means 21 for collecting the amount of resources allocated to the guest computer 22, and a resource allocation indicating the resource allocation amount for this guest computer 22 and its set time. And a resource allocation information storage unit 23 for sequentially storing information.

  The guest computer 22 is a virtual computer constructed in the host computer 2 and has a function of acting as a computer that executes arithmetic processing using computer resources allocated from the host computer 2.

As shown in FIG. 1, the guest computer 22 includes a performance measurement unit 221 that measures the load on the guest computer 22 and a performance information storage unit 222 that stores performance information indicating the measured load and measurement time. ing. Here, the load measured by the performance measuring unit 221 includes the CPU usage rate per unit time of the guest computer 22 and the number of job processes per unit time.

  The virtual machine management system 1 includes an information input unit 11 that inputs a correction command in accordance with an operation of the administrator, a performance information acquisition unit 12 that acquires performance information indicating the load on the guest computer 22 from the guest computer 22 side, and an information input unit. 11 is provided with resource allocation information acquisition means 13 for acquiring, from the host computer 2 side, resource allocation information indicating the amount of computer resources allocated to the guest computer 22 in accordance with the correction command input to 11.

  Further, the virtual machine management system 1 corrects the performance information of the guest computer 22 acquired by the performance information acquisition unit 12 based on the resource allocation information acquired by the resource allocation information acquisition unit 13 and corrects the corrected information. A performance information correction unit 14 that outputs data for monitoring for monitoring the guest computer 22 is provided.

  Further, the virtual machine management system 1 receives the monitoring data output from the performance information correction unit 14 and generates graph data which is image data of a monitoring graph representing the history of the monitoring data; And a display unit 16 for displaying a graph representing a load variation of the guest computer 22 based on the graph data generated by the graph generation unit 15.

The information input unit 11 sends a correction command input according to the operation of the administrator to the resource allocation information acquisition unit 13. This correction command specifies a target period for graph display. Information input means 1 1, as a means for accepting an operation of the administrator, for example, a keyboard, an input device such as a mouse.

  The performance information acquisition unit 12 has a function of reading out and acquiring the performance information of the guest computer 22 from the performance information storage unit 222 at a constant period, and outputting it to the performance information correction unit 14. Here, the performance information acquired by the performance information acquisition unit 12 includes the CPU usage rate of the guest computer 22 and the number of job processes per unit time.

  The resource allocation information acquisition unit 13 has a function of reading and acquiring the history of resource allocation information for the period specified by the correction command from the information input unit 11 from the resource allocation information storage unit 23 and outputting the history to the performance information correction unit 14. I have. Here, as the resource allocation amount acquired by the resource allocation information acquiring unit 13, various things such as the number of guest computers 22 on the host computer 2, the CPU clock and the number of CPUs allocated to the guest computer 22 can be used. It is.

  The performance information correction unit 14 inputs the performance information output from the performance information acquisition unit 12 and the resource allocation information output from the resource allocation information acquisition unit 13, and sets the load value of the guest computer 22 indicated by the performance information. On the other hand, by multiplying the adjustment factor corresponding to the amount of resources allocated to the guest computer 22 at the time of the measurement, the performance information for the period specified by the correction command is corrected, and the corrected information is graphed as monitoring data. A function for outputting to the generation means is provided.

  The adjustment magnification is a magnification set in advance to cancel the influence of the resource allocation amount to the guest computer 22 on the load measurement value of the guest computer 22 and varies depending on the resource allocation amount to the guest computer 22. Value. A value obtained by multiplying the load value of the guest computer 22 by the adjustment rate according to the resource allocation amount at that time is a value excluding increase / decrease due to the setting change of the resource allocation amount.

  Since the correspondence between the amount of resources allocated to the guest computer 22 and the adjustment factor is determined by the performance of the host computer 2, the adjustment factor is changed while changing the resource allocation amount to the guest computer 22 during system development. It may be set in advance based on data obtained by repeating a test for measuring the load.

  Here, the function of the performance information correction unit 14 will be described with reference to FIG. 2, taking as an example the case where the load indicated by the performance information acquired by the performance information acquisition unit 12 is the CPU usage rate of the guest computer 22.

FIG. 2A is a graph showing an example of the CPU usage rate history of the guest computer 22 corrected by the performance information correcting unit 14. FIG. 2B is a graph showing an example of the CPU usage rate history of the guest computer 22 indicated by the performance information acquired by the performance information acquisition unit 12. FIG. 2 ( c ) is a graph showing an example of the history of the CPU allocation rate for the guest computer 22.

  As shown in FIG. 2C, the CPU allocation rate input from the resource allocation information acquisition unit 13 by the performance information correction unit 14 is set to 50% between times t0 and t1 and 25% between times t1 and t2. If it has been done, as shown in FIG. 2B, the CPU usage rate of the guest computer 22 input from the performance information acquisition unit 12 by the performance information correction unit 14 is increased between times t1 and t2.

  This is not a change in the processing executed by the guest computer 22 at time t1, but an increase due to a change in the scale of the measurement value due to a change in the allocation of CPU resources. Therefore, such an increase in load returns to normal due to a change in the resource allocation setting for the guest computer 22 on the host computer 2, and therefore it is not necessary to present the increase in load during this period to the administrator as a failure. .

  Therefore, the performance information correcting unit 14 multiplies the CPU usage rate of the guest computer 22 by an adjustment factor corresponding to the CPU allocation rate for the guest computer 22 in order to cancel the CPU usage rate variation due to the CPU allocation rate setting change. Then, the values are corrected to the values shown in FIG. Here, the correspondence relationship between the CPU allocation rate and the adjustment magnification may be expressed by a curve shown in FIG. 3, and the performance information correction unit 14 has a mathematical expression indicating the curve shown in FIG. The CPU allocation rate input from 13 is substituted into the equation to calculate the corresponding adjustment magnification, and this calculated adjustment magnification is multiplied by the CPU allocation rate. The correspondence between the CPU allocation rate and the adjustment magnification varies depending on the capability of the host computer 2.

  Since this performance information correction means 14 can cancel the increase or decrease in the load amount of the guest computer 22 due to the setting change of the resource allocation amount for the guest computer 22, the corrected data is used as monitoring data, and the failure of the guest computer 22 By using it for detection, it is possible to reduce erroneous detection of a fault.

  Here, the performance information correction unit 14 is limited to a configuration in which the monitoring data is calculated by multiplying the load value of the guest computer 22 by an adjustment factor corresponding to the amount of resources allocated to the guest computer 22 at the time of measurement. First, the resource amount allocated to the guest computer 22 at the reference time specified in the correction command is determined as the reference value, and the adjustment magnification for the load value is set only when the resource allocation amount is set to a value different from the reference value. You may comprise so that it may apply. The adjustment magnification in this case may be a value obtained by dividing the adjustment magnification corresponding to the load value to be corrected calculated based on the mathematical expression of the curve shown in FIG. 3 by the adjustment magnification corresponding to the reference value obtained from the mathematical expression. That's fine.

  The graph generation unit 15 sequentially inputs the monitoring data output from the performance information correction unit 14, generates graph data that is image data of the monitoring graph representing the history of the sequentially input monitoring data, and displays the data. 16 is output. In the case described with reference to FIG. 2, graph data that is image data of the graph shown in FIG.

  The graph generation unit 15 further inputs the resource allocation information from the resource allocation information acquisition unit 13 and compresses or reduces the size of the time axis scale interval in the monitoring graph based on the resource allocation amount indicated by the resource allocation information. Has the ability to expand.

  Here, the compression or expansion function of the time axis scale interval in the graph generation unit 15 will be described with reference to FIG. 4, taking as an example the case where the load indicated by the performance information is the number of job processes per unit time of the guest computer 22. To do.

  FIG. 4A is a graph showing the history of the number of job processes of the guest computer 22 corrected by the performance information correction unit 14. FIG. 4B is a graph showing the history of the number of job processes of the guest computer 22 acquired by the performance information acquisition unit 12. FIG. 4C is a graph showing the history of the CPU allocation rate for the guest computer 22 acquired by the resource allocation information acquisition unit 13.

  As shown in FIG. 4C, when the CPU allocation set to 50% is set to 25% between times t3 and t4, the CPU performance of the guest computer 22 decreases during times t3 to t4. As shown in FIG. 4B, the number of job processes measured by the guest computer 22 decreases. As shown in FIG. 4B, the job information per unit time is obtained simply by the graph generation unit 15 graphing the data corrected by the performance information correction unit 14 with respect to the job processing number of the guest computer 22 as shown in FIG. Since the number appears to increase, the graph generation unit 15 compresses or expands the size of the scale interval on the time axis of the graph based on the CPU allocation rate to create graph data (FIG. 4A )reference). As a result, it is possible to generate a graph that excludes fluctuations in the number of job processes due to fluctuations in CPU allocation, while accurately representing the number of job processes per unit time.

  As described above, the virtual machine management system 1 according to the present embodiment multiplies the performance information acquired by the performance information acquisition unit 12 by an adjustment factor corresponding to the resource allocation information acquired by the resource allocation information acquisition unit 13. Therefore, the performance information correction means 14 for correcting the performance information and outputting it as monitoring data is provided, so that the load on the guest computer 22 is represented without being affected by the setting change of the resource allocation amount to the guest computer 22. Can be obtained as monitoring data.

  Since this monitoring data does not increase or decrease rapidly even if the resource allocation amount to the guest computer 22 is changed, it is possible to set a failure determination threshold for this monitoring data. Therefore, it is possible to easily carry out accurate detection of a failure occurring in the system.

  Furthermore, the virtual machine management system 1 according to the present embodiment sequentially inputs monitoring data from the performance information correction unit 14 and generates graph data which is image data of a monitoring graph representing the history of the monitoring data. And a display unit 16 for displaying a graph representing a change in the load of the guest computer 22 based on the graph data generated by the graph generation unit 15, the graph displayed on the display unit 16 is the guest computer. No change in the load amount of the guest computer 22 due to the setting change of the resource allocation amount with respect to 22 is shown, and the administrator accurately determines the presence or absence of a failure occurring in the guest computer 22 based on the graph displayed on the display means 16. Can be determined.

  Further, the virtual machine management system 1 of the present embodiment is configured to multiply the load measured by the guest machine 22 by an adjustment rate according to the resource allocation amount. It can be applied to all 22 load items.

  Next, the operation of the virtual machine management system 1 of this embodiment will be described. Here, the following operation description is an embodiment of the virtual machine management method of the present invention.

FIG. 5 is a flowchart showing the operation of the virtual machine management system 1 of this embodiment. As shown in FIG. 5, in the virtual machine management system 1 of the present embodiment, first, the performance information acquisition unit 12 acquires the performance information of the guest computer 22 from the performance information storage unit 222 and performs the performance information correction unit 14. (Step S901 in FIG. 5).

When a correction command is input to the information input unit 11, the input correction command is sent to the resource allocation information acquisition unit 13 (Yes in step S902 in FIG. 5). Then, the resource allocation information acquisition unit 13 that has input the correction command acquires the resource allocation information related to the guest computer 22 for the specified period from the resource allocation information storage unit 23 and outputs it to the performance information correction unit 14 (FIG. Step 903).

Then, the performance information correction unit 14, to the load value of the guest computer 22 indicated by the performance information from the performance information acquisition unit 12, assigned to a guest computer 2 2 indicated resource allocation information from resource assignment information obtaining unit 13 The performance information is corrected by multiplying by the adjustment magnification corresponding to the amount of resources, and the corrected information is output as monitoring data to the graph generating means (step 904 in FIG. 5).

  The graph generation unit 15 to which the monitoring data is input generates graph data of the monitoring graph representing the history of the monitoring data and sends it to the display unit 16 (step 905 in FIG. 5). At this time, the graph generation unit 15 inputs the resource allocation information from the resource allocation information acquisition unit 13 and compresses the size of the time axis scale interval in the monitoring graph based on the resource allocation amount indicated by the resource allocation information. Or it may be expanded.

  Then, the display means 16 to which the graph data is input displays a graph representing the load fluctuation of the guest computer 22 based on this graph data (step 906 in FIG. 5). An example of the graph displayed by the display means 16 is as shown in FIG. 2 (a) or FIG. 4 (a).

  As described above, when displaying the performance information of the guest computer 22, the virtual computer management system 1 of the present embodiment performs measurement by the guest computer 22 based on the setting status of resource allocation for the guest computer 22 on the host computer 2. Corrected performance information. Therefore, it is possible to display a graph representing the load on the guest computer 22 while eliminating the influence of the change of the resource allocation amount on the guest computer 22.

  Here, the above-described performance information acquisition unit 12, resource allocation information acquisition unit 13, performance information correction unit 14, and graph generation unit 15 may be configured such that their function contents are programmed and executed by a computer. .

  As described above, the virtual computer management system 1 according to the present embodiment corrects the performance information of the guest computer 22 according to the setting status of resource allocation for the guest computer 22 on the host computer 2. The influence of the resource allocation amount on the performance information of the guest computer 22 can be eliminated, and the load caused by the internal processing of the guest computer 22 can be accurately indicated. As a result, it is possible to observe the exact fluctuation of the load on the guest computer 22, to effectively suppress erroneous detection and detection omission of a failure occurring in the guest computer 22, and to improve the efficiency of management work for the guest computer 22. it can.

  Further, the virtual machine management system 1 according to the present embodiment simply multiplies the performance information measured by the guest machine 22 by an adjustment factor according to the amount of resources allocated to the guest machine 22 at the time of measurement. Since the correction is performed, not only the CPU usage rate but also any load item that varies depending on the resource allocation amount such as the number of job processes can be adopted as the load item to be monitored.

  The present invention can be applied to a performance information monitoring application of a virtual machine for a system that dynamically changes resource allocation to a virtual machine.

DESCRIPTION OF SYMBOLS 1 Virtual computer management system 2 Host computer 11 Information input means 12 Performance information acquisition means 13 Resource allocation information acquisition means 14 Performance information correction means 15 Graph generation means 16 Display means 21 Resource allocation collection means 22 Guest computer (virtual computer)
23 Resource Allocation Information Storage Unit 221 Performance Measuring Means 222 Performance Information Storage Unit

Claims (6)

A virtual machine management system that monitors the operating status of a guest computer that is a virtual machine to which computer resources are dynamically allocated by a host computer,
Performance information acquisition means for acquiring performance information indicating the load of the guest computer;
Resource allocation information acquisition means for acquiring resource allocation information indicating the amount of computer resources allocated in advance to the guest computer from the host computer side;
A performance information correction unit that corrects the performance information acquired by the performance information acquisition unit based on the resource allocation information acquired by the resource allocation information acquisition unit, and outputs the corrected information as monitoring data ;
Graph generating means for inputting the monitoring data from the performance information correcting means and generating graph data that is image data of a monitoring graph representing a history of the monitoring data;
The graph generation unit inputs resource allocation information from the resource allocation information acquisition unit, and based on the resource allocation amount to the guest computer indicated by the resource allocation information, the size of the time interval on the time axis in the monitoring graph A virtual machine management system having a function of compressing or expanding the image data and converting the image data of the graph into the graph data . In the virtual machine management system according to claim 1,
The performance information correction means multiplies an adjustment factor corresponding to a resource allocation amount indicated by the resource allocation information by a load value indicated by the performance information, and uses the calculated information as the monitoring data. Virtual computer management system. In the virtual machine management system according to claim 1 or 2,
A virtual computer management system further comprising display means for displaying a graph representing a load fluctuation of the guest computer based on the graph data generated by the graph generation means . In the virtual machine management system according to any one of claims 1 to 3 ,
The virtual machine management system characterized in that the performance information acquired by the performance information acquisition means is information indicating a CPU usage rate or job processing count of the guest computer. The performance information acquisition means acquires performance information indicating the load of the guest computer, which is a virtual computer to which computer resources are dynamically allocated by the host computer,
Before and after the performance information acquisition, resource allocation information acquisition means acquires resource allocation information indicating the amount of computer resources allocated in advance to the guest computer from the host computer side ,
The performance information performance information correction hand stage is corrected by multiplying the adjustment magnification factor corresponding to the resource allocation amount of the obtained resource allocation information indicates the load value that the indicated acquired performance information, the information after the correction Output as monitoring data for monitoring the guest computer ,
Based on the resource allocation amount indicated by the resource allocation information, the graph generation means compresses or expands the size of the time axis scale interval in the monitoring graph representing the history of the monitoring data, and is image data of this graph. A virtual computer management method , wherein the graph generation means generates graph data . A performance information acquisition function for acquiring performance information indicating a load of a guest computer that is a virtual computer to which computer resources are dynamically allocated by a host computer;
A resource allocation information acquisition function for acquiring resource allocation information indicating the amount of computer resources allocated in advance to the guest computer;
The adjustment information corresponding to the resource allocation amount indicated by the acquired resource allocation information is multiplied by the load value indicated by the acquired performance information to correct the performance information, and the corrected information is monitored by the guest computer. Performance information correction function to output as monitoring data for
Based on the resource allocation amount indicated by the resource allocation information, the scale interval on the time axis in the monitoring graph representing the history of the monitoring data is compressed or expanded, and the graph data which is image data of the graph is A virtual machine management program characterized by causing a computer to realize a graph generation function to be generated .

Images