JP2013206368A - Virtual environment operation support system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To efficiently execute a failure detection caused by a physical device shared over physical machines in a virtual environment in which a configuration change in a virtual machine can occur.SOLUTION: The virtual environment operation support system relating to this invention records a history of use relations between a physical device shared by a plurality of physical machines and respective virtual machines, and specifies a use state of the physical device of each virtual machine in a plurality of dates and hours on the basis of the history.

Description

  The present invention relates to operation of a virtual environment system, and relates to a technology that supports efficient monitoring of virtual machine performance and quick failure handling.

  The use of virtualization technology is progressing for the purpose of effective utilization of IT resources. In a system environment (virtual environment) using virtualization technology, one physical machine can be separated by software technology and a plurality of virtual machines can be operated. By replacing a server group that has conventionally been operated by a plurality of physical machines with a virtual machine and consolidating them on one physical server, it is possible to improve the utilization efficiency of IT resources.

  On the other hand, operations of virtual machines in a virtual environment, particularly virtual machine performance management and troubleshooting, have problems unique to the virtual environment different from the conventional physical machine environment.

  FIG. 1 is a diagram illustrating an example of a failure unique to a virtual environment. Here, a decrease in virtual machine performance due to a change in the usage status of a storage device that is a resource shared by each physical machine is shown. The virtual machine a and the virtual machine b in FIG. 1 operate while sharing the physical machine A and the storage 1. In this environment, when the virtual machine a consumes a large amount of resources such as CPU, memory, and storage I / O, the virtual machine b sharing resources with the virtual machine a is sufficient from the physical machine A and the storage 1. The amount of resources is not allocated, and the performance of the virtual machine b is degraded. Furthermore, since the storage 1 also uses the virtual machine c and the virtual machine d via the physical machine B, the storage I / O performance of the virtual machine c and the virtual machine d also deteriorates.

  The above-mentioned degradation of virtual machine performance due to changes in the use of shared resources is difficult to find and identify with conventional machine-level performance monitoring. Performance monitoring combined with comprehension is essential.

  For example, in the example of FIG. 1, even if a performance monitoring tool is incorporated in each virtual machine and the performance degradation of the virtual machine b can be detected, the investigation of only the virtual machine b that is the detection target is the root cause of the performance degradation. Since (a large amount of shared resources of the virtual machine a) cannot be found, it is difficult to deal with it quickly. In order to quickly deal with failures, the dependency relationships of system components such as virtual machines, physical machines, storage, and network devices are always recorded, and when failures such as degradation of virtual machine performance occur, they are recorded. It is effective to analyze the cause of failure including the use status of shared resources based on the system configuration information. In particular, in a large-scale virtual environment where the system configuration is complicated and the frequency of system configuration change is high, efficient grasping and utilization of system configuration information is important for performance management of virtual machines.

  Examples of existing technologies that have a system configuration information recording function in a virtual environment and can be used for performance management and troubleshooting of virtual machines include the following.

  In Patent Document 1, a dependency rule for identifying the cause of a failure is created based on business / device / network configuration information input in advance by an administrator. There has been disclosed a technique for presenting a failure cause candidate to a system administrator by matching failure site information communicated by an event with a failure identification rule created in advance.

  Patent Document 2 includes a physical device in which a failure has occurred and an affected range specifying table that stores a range of logical devices affected by the physical device, and is notified by the failed event when the failed event is notified. A technique for extracting a relationship between a physical device and a virtual device that is affected by a failure by referring to a failure impact range table based on the failed site information and presenting a failure impact range and a failure cause candidate is disclosed.

JP 2011-113122 A JP 2010-9411 A

  The above-described conventional techniques have the following problems, and it is difficult to use them for performance management and troubleshooting of virtual machines in a large-scale virtual environment.

  In the technique described in Patent Document 1, the administrator has to manually input system configuration information necessary for system operation. The virtual environment has a higher frequency of configuration changes than the physical environment. This is because in a virtual environment, the placement of virtual machines is frequently changed in accordance with the resource usage status in order to improve the utilization efficiency of shared resources such as physical machines and storage. Some virtualization software intended for use in a large-scale environment has a function for automatically executing such virtual machine relocation. It is very difficult for the administrator to manually update the system configuration information every time the virtual machine is moved.

  In the technique described in Patent Document 2, since the creation of system configuration information is automated, application to a large-scale virtual environment is easier than in Patent Document 1. On the other hand, the system configuration information that is automatically generated always keeps only the latest state, so when investigating the root cause of virtual machine performance degradation, the current system configuration where performance degradation occurred and the time when normal operation was performed in the past It is difficult to investigate the cause of the failure by matching with the system configuration.

  The present invention has been made in view of the above problems, and in a virtual environment in which a configuration change of a virtual machine can occur, efficiently detects a failure caused by a physical device shared across physical machines. The purpose is to do.

  The virtual environment operation support system according to the present invention records a history of usage relations between a physical device shared by a plurality of physical machines and each virtual machine, and based on the history of each virtual machine at a plurality of dates and times. Identify the usage status of physical devices.

  According to the virtual environment operation support system of the present invention, it is possible to efficiently detect a failure that occurs across physical machines and physical devices even in a virtual environment in which a configuration change of a virtual machine occurs.

It is a figure which shows the example of a failure peculiar to virtual environment. 1 is a diagram illustrating a schematic configuration of a virtual environment operation support system 1000 according to Embodiment 1. FIG. 2 is a diagram illustrating a configuration of a virtual machine performance management server 202. FIG. It is a figure which shows the structure of the resource usage record storage table. It is a figure which shows the structure of the structure information storage table 314. FIG. It is a figure which shows the structure of the structure change event storage table 315. FIG. It is an image figure of the virtual machine performance analysis screen 701. It is an image figure of the performance degradation cause analysis screen. It is a conceptual diagram which shows the whole processing flow of the virtual environment operation support system 1000. It is a figure explaining the processing flow of the data collection program. It is a figure explaining the processing flow of the virtual machine performance analysis screen creation program. It is a figure explaining the processing flow of the performance fall cause analysis screen creation program. It is a figure explaining the processing flow of the resource use performance analysis program. It is a figure explaining the processing flow of the system configuration information analysis program.

  Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. However, it should be noted that the embodiment described below is merely an example for realizing the present invention and does not limit the technical scope of the present invention.

<Embodiment 1>
In the first embodiment of the present invention, a storage apparatus will be described as an example of a resource shared across physical machines. It is assumed that the storage device is always operating and the state does not change. It is assumed that the same name does not exist for the storage name, physical machine name, and virtual machine name, and the target can be uniquely specified.

<Embodiment 1: System configuration>
FIG. 2 is a diagram showing a schematic configuration of the virtual environment operation support system 1000 according to the first embodiment of the present invention. The virtual environment operation support system 1000 has a configuration in which a virtual environment management server 201, a virtual machine performance management server 202, a virtual machine performance management client 203, and a plurality of physical machines 204 are connected via a management LAN (Local Area Network) 205. Take. The physical machine 204 is connected to the storage 206 via the storage network 207.

  The physical machine 204 is a computer that operates the virtualization software 208 and the virtual machine 209. The virtualization software 208 is software that emulates the hardware configuration of the physical machine 204 using software technology and enables the virtual machine 209 having an arbitrary hardware configuration to operate.

  The virtual environment management server 201 is a computer that operates the virtual environment management software 210. The virtual environment management software 210 collects information related to the physical machine 204 or virtual machine 209 in which the virtualization software 208 is installed via the management LAN 205 and temporarily stores the information, and provides the stored information to external software. It has a function of providing a virtual environment management API (Application Program Interface).

  The virtual machine performance management server 202 is a computer that executes the main processing of the present invention. The virtual machine performance management server 202 is a data collection program 211, a resource usage record analysis program 212, a system configuration information analysis program 213, and other sub-details described in FIG. Stores programs and table data, and has functions to execute and manage them.

  The virtual machine performance management client 203 is a computer operated by a user of the virtual environment operation support system 1000. The virtual machine performance management client 203 includes at least an input unit that receives user input information, an output unit such as a display that presents information to the user, and a communication processing unit that processes communication via the management LAN. In response to a user system usage request received from the input unit, the virtual machine performance management server 202 is accessed and screen data created by the virtual machine performance management server 202 is displayed on the output unit.

  In the virtual environment, the physical machine 204 is usually connected to a service providing LAN that is different from the management LAN 205 for the virtual machine 209 to communicate with the external network. The explanation is omitted because it is not involved.

<Embodiment 1: Configuration of virtual machine performance management server>
FIG. 3 is a diagram illustrating the configuration of the virtual machine performance management server 202. Note that the configurations of the virtual environment management server 201 and the virtual machine performance management client 203 are the same as those in FIG. 3 except for the storage information in the auxiliary storage unit 306 described later, and thus description thereof is omitted.

  The virtual machine performance management server 202 loads an input unit 301 for inputting commands and data, an output unit 302 for outputting system status, an arithmetic unit 303 for executing and calculating programs, and loading programs and data. The main storage unit 304, a communication processing unit 305 that establishes a connection with another communication device, and an auxiliary storage unit 306 that stores data and programs are connected via a bus 307.

  The auxiliary storage unit 306 includes a data collection program 211, a virtual machine performance analysis screen creation program 309, a performance degradation cause analysis screen creation program 310, a resource usage record analysis program 212, a system configuration information analysis program 213, a resource A usage record storage table 313, a configuration information storage table 314, and a configuration change event storage table 315 are stored.

  The data collection program 211 is a program that collects resource usage record information and system configuration information of the virtual machine 209, the physical machine 204, and the storage 206 using the virtual environment management API provided by the virtual environment management software 210. The resource specification record information is information describing a record of the use of the storage 206 by the virtual machine 209 or the physical machine 204. The system configuration information is information describing a connection relationship between the virtual machine 209 or the physical machine 204 and the storage 206. That is, the information describes the situation in which the virtual machine 209 or the physical machine 204 is using the storage 206.

  The virtual machine performance analysis screen creation program 309 is a program for creating the virtual machine performance analysis screen shown in FIG. In the course of processing, the resource use record analysis program 212 and the resource use record information storage table 313 are used.

  The performance degradation cause analysis screen creation program 310 is a program that creates the performance degradation cause analysis screen shown in FIG. In the course of processing, the system configuration information analysis program 213, the resource usage record information storage table 313, the configuration information storage table 314, and the configuration change event storage table 315 are used.

  Hereinafter, for convenience of description, each program may be described as an operation subject. However, it is added that a function unit having an arithmetic function such as the arithmetic unit 303 actually executes these programs.

<Embodiment 1: Resource usage record storage table>
FIG. 4 is a diagram showing the configuration of the resource usage record storage table 313. As shown in FIG. The resource usage record storage table 313 is a table for storing resource use record information of virtual machines, physical machines, and storages collected by the virtual machine performance management server 202 by executing the data collection program 211.

  The information collection date 3131 holds the collection date of the resource usage record information to be stored. The entity name 3132 holds information for uniquely identifying the component in the virtual environment of the resource usage record information to be stored. In the first embodiment, it is one of a virtual machine name, a physical machine name, and a storage name. The resource type 3133 holds the resource type of the resource usage record information to be stored. In the first embodiment, the values related to the I / O of the storage 206 are targeted. However, the present invention is not limited to this, and the CPU usage rate, the memory usage amount, the network bandwidth usage rate, and the like can also be handled. The result value 3134 holds the performance value of the resource use result information specified by the value of the resource type 3133.

<Embodiment 1: Configuration information storage table>
FIG. 5 is a diagram showing the configuration of the configuration information storage table 314. The configuration information storage table 314 is a table that stores system configuration information of a virtual environment at a certain point collected by the virtual machine performance management server 202 by executing the data collection program 211.

  The information collection date 3141 holds the collection date of system configuration information to be stored. The storage name 3142, the physical machine name 3143, and the virtual machine name 3144 respectively hold the storage name, physical machine name, and virtual machine name that are the acquisition targets of the stored system configuration information. The virtual machine state 3145 and the physical machine state 3146 hold state information at the time of acquisition of the virtual machine and physical machine specified by the virtual machine name 3144 and the physical machine name 3143, respectively. The status information includes, for example, power on, power off, hibernation.

<Embodiment 1: Configuration change event storage table>
FIG. 6 is a diagram showing the configuration of the configuration change event storage table 315. The configuration change event storage table 315 is a table that stores event information that is recorded by the virtual machine performance management server 202 by executing the data collection program 211 and records changes in the system configuration and state of the virtual environment. For example, event information such as the movement of the virtual machine 209 to another physical machine 204, the power on / off of the virtual machine 209 or the physical machine 204 being executed, or the transition to the hibernation state / resumption has been generated. Recorded with date and time.

  The event occurrence date and time 3151 holds the collection date and time of the event information to be stored. The entity name 3152 holds information for uniquely specifying the component in the virtual environment of the event information to be stored. In the first embodiment, it is either a virtual machine name or a physical machine name. The generated event name 3153 holds an event name indicating the type of event to be stored. The migration source entity name 3154 and the migration destination entity name 3155 hold the migration source and migration destination entity names when the virtual machine 209 or the physical machine 204 moves to another physical machine 204 or storage 206, respectively. Note that the values of these fields are the same for events that do not involve movement.

<Embodiment 1: Virtual machine performance analysis screen>
FIG. 7 is an image diagram of the virtual machine performance analysis screen 701. The virtual machine performance analysis screen 701 is a screen used for efficiently monitoring the performance state of a large number of virtual machines 209 and quickly identifying a virtual machine whose performance is degraded. The virtual machine performance analysis screen 701 is created by executing the virtual machine performance analysis screen creation program 309.

  The virtual machine performance analysis screen 701 has a virtual machine performance trend display area 702, a performance degradation virtual machine name display area 703, a data update button 704, and a performance analysis button 705. The virtual machine performance trend display area 702 has a detection threshold value input field 706 and a detection threshold value sign 707. The performance degradation virtual machine name display area 703 has a performance degradation virtual machine name button 708.

  The virtual machine performance trend display area 702 is an area for displaying statistical information on performance trends of virtual machines. By referring to this area, the user can confirm the presence of a virtual machine whose performance is degraded (performance degraded virtual machine). In the first embodiment, as an example of statistical information, the average of performance results for each virtual machine and each resource type over a certain period is aggregated (average CPU allocation waiting time, average memory swap amount, average network packet loss rate, Average storage I / O processing wait time), and the frequency distribution diagram is displayed with the horizontal axis (data range) as the average value and the vertical axis (frequency) as the number of virtual machines that fall within each data range. .

  A detection threshold value input field 706, a detection threshold sign 707, and a performance analysis button 705 are used to specify extraction conditions for virtual machines displayed in the performance degradation virtual machine name display area 703. While confirming the statistical information displayed in the virtual machine performance trend confirmation area 702, the user inputs a performance value condition to be extracted as a virtual machine whose performance is degraded in the detection threshold input field 706. When a value is input to the detection threshold value input field 706, the detection threshold sign 707 is redrawn so as to indicate the position of the detection threshold value on the statistical information. When the user presses the performance analysis button 705 after inputting the detection threshold value input field 706, the virtual machine name extracted according to the specified condition is displayed in the performance degradation virtual machine name display area 703.

  The degraded performance virtual machine name display area 703 is an area for displaying the virtual machine name extracted according to the conditions specified in the virtual machine performance trend confirmation area 702. By referring to this area, the user can specify a virtual machine whose performance is degraded from a large number of virtual machines. Note that the virtual machine name for performance degradation is displayed after ranking according to the degree of performance degradation in order to make it easy to confirm the severity of performance degradation. For example, in the first embodiment, virtual machines with a larger number of resource types that exceed the detection threshold are displayed at the top. If the number of resource types exceeding the detection threshold is the same, the higher the degree of exceeding, the higher the display.

  The virtual machine name displayed in the degraded performance virtual machine name display area 703 is presented as a degraded performance virtual machine name button 708. The performance degradation virtual machine name button 708 is a button in which the performance degradation virtual machine name is described. When pressed, the screen is changed to a performance degradation cause analysis screen (FIG. 8) for investigating the cause of performance degradation of the virtual machine. Button.

  When the data update button 704 is pressed, the data collection program 211 is executed, the latest information is acquired from the virtual environment management server 201, and the virtual machine performance analysis screen is redrawn.

  The detection threshold value input field 706 can be automatically designated by a method such as automatically setting performance values corresponding to the top 10% in order from the lowest performance value in the distribution of performance values.

<Embodiment 1: Performance degradation cause analysis screen>
FIG. 8 is an image diagram of the performance degradation cause analysis screen 801. The performance degradation cause analysis screen 801 is a screen used for confirming a situation where the performance degradation virtual machine specified on the virtual machine performance analysis screen 701 uses a shared resource and investigating the cause of performance degradation. The performance degradation cause analysis screen 801 is created by executing the performance degradation cause analysis screen creation program 310.

  The performance degradation cause analysis screen 801 includes a virtual machine performance time series display area 802, a shared resource usage status display area 803, a normal operation date / time input field 804, a resource type input field 805, a configuration display button 806, and a return button. 807. The virtual machine performance time series display area 802 has a virtual machine name display field 808 and a normal operation date / time sign 809. The shared resource usage status display area 803 has a normal operation date / time shared resource usage status display column 810 and a current date / time shared resource usage status display column 811. The normal operation date / time shared resource usage status display column 810 and the current date / time shared resource usage status display column 811 are configured by a system configuration map 812, a resource status sign 813, and a resource usage amount display sign 814.

  The virtual machine performance time series display area 802 is an area for displaying time series data of the performance value of the designated virtual machine. By referring to this area, the user can confirm the fluctuation of the virtual machine performance value and specify the date and time when the performance degradation occurred. In the first embodiment, the time-series changes in the performance performance values for each virtual machine and each resource type in a certain period, which are the same as those statistically displayed in the virtual machine performance trend display area 702, are displayed in a graph. In the virtual machine name display field 808, the name of the virtual machine displaying the performance time series is displayed.

  The shared resource usage status display area 803 is an area for displaying the shared resource usage status when the specified performance-reduced virtual machine is operating normally and the shared resource usage status when there is an abnormality. For normal operation, the user designates and inputs the date and time when the virtual machine is estimated to have been operating normally in the normal operation date and time input field 804 described later. When there is an abnormality, the user may specify and input, or the current date and time may be automatically specified instead. By referring to this area, the user can compare the shared resources, the status of each resource, and the load status of each resource when the virtual machine with degraded performance is operating normally and abnormally. It is possible to efficiently investigate and identify the cause of the decrease in

  A normal operation date / time input field 804, a normal operation date / time sign 809, a resource type input field 805, and a configuration display button 806 are used to specify conditions for the shared resource usage status displayed in the shared resource usage status display area 803. The user refers to the time series of performance values displayed in the virtual machine performance time series display area 802, identifies the date and time when the virtual machine was operating normally, and enters this date and time in the normal operation date and time input field 804. input. As the normal operation date and time, for example, the date and time before the performance record value shows deterioration may be specified. When the date and time is entered in the normal operation date and time input field 804, the normal operation date and time display sign 809 is redrawn so as to indicate the position of the normal operation date and time on the time series. Further, the user inputs the resource type to be compared in the shared resource usage status display area 803 in the resource type input field 805 and presses the configuration display button 806. When the configuration display button 806 is pressed, the shared resource usage status display area 803 displays the usage status of the shared resource at the date and time entered in the normal operation date and time input field 804 and the sharing at the current time (or when an abnormality is specified separately). The resource usage status is displayed in the normal operation date and time shared resource usage status display column 810 and the current date and time shared resource usage status display column 811, respectively.

  The system configuration map 812 shows the logical configuration of the shared resource at a specified time. By referring to the system configuration map 812, the user can obtain information regarding candidates for causes of performance degradation such as addition of virtual machines that did not exist before performance degradation.

  The resource status sign 813 indicates the status of the shared resource at a specified time. The state includes, for example, power on / off and hibernation. By referring to the resource status sign 813, the user can obtain information on candidates for the cause of performance degradation, such as activation of a virtual machine that was stopped before the performance degradation.

The resource usage amount display sign 814 displays the usage amount of the resource designated in the resource type input field 805 in a stepwise manner. By referring to the resource usage amount display sign 814, the user can obtain information on candidates for the cause of performance degradation, such as a sudden increase in the shared resource usage load of the virtual machine that had a low load before the performance degradation.
A return button 807 is a button for ending the display of the performance degradation cause analysis screen 801 and transitioning to the virtual machine performance analysis screen 701.

<Embodiment 1: Initial conditions for explaining processing flow>
Hereinafter, a processing flow of the virtual environment operation support system 1000 according to the first embodiment of the present invention will be described. However, it is assumed that the following initial values are set in the following description. The initial value can be changed by a system setting parameter or the like.

  The statistical information aggregation period in the virtual machine performance trend display area 702 in FIG. 7 is the period from the execution time of the virtual machine performance analysis screen creation program 309 to the past one hour, and the performance value type to be aggregated is CPU latency / memory Swap amount, network packet loss rate, storage I / O command processing wait time. The detection threshold 706 takes a null character when nothing is specified.

  The initial values of the time series display period and performance value type in the virtual machine performance time series display area 802 in FIG. 8 are the same period and performance value type as the statistical information in the virtual machine performance trend display area 702. If nothing is designated as the normal operation date and time 804, for example, the normal operation date and time 804 is set to the same time three days before the execution date and time of the performance degradation cause analysis screen creation program 310. For the resource type 805, the CPU usage rate, the memory usage rate, the network communication amount, and the storage I / O amount are registered as selection items in advance. If nothing is specified, for example, the storage I / O amount is specified. Shall.

  The threshold for displaying the resource usage amount display sign 814 in stages is assumed to be specified in advance for each resource type. For example, in the case of a CPU usage rate, a threshold value such as less than 30%, 30% or more and less than 60%, or 60% or more is designated. Or, the CPU usage rate is aggregated for all virtual machines, and the threshold generation rule based on the virtual machine ranking is set such that the top 30% of the virtual machines with the highest usage rate are level 3 and the middle 30% are level 3. You may register.

  In the following description, the virtual machine performance analysis screen creation program 309 and the data collection program 211 are periodically executed, but this execution cycle is specified in advance. For example, the execution cycle of the virtual machine performance analysis screen creation program 309 is specified as 1 hour, and the execution cycle of the data collection program 211 is specified as 24 hours.

<Embodiment 1: Overview of overall processing>
FIG. 9 is a conceptual diagram showing an overall processing flow of the virtual environment operation support system 1000. The virtual environment operation support system 1000 is activated when a user operates the virtual machine performance management client 203 to access the virtual machine performance management server 202 and request a virtual machine performance analysis screen. In response to this request, the virtual machine performance management server 202 executes the virtual machine performance analysis screen creation program 309.

  The virtual machine performance analysis screen creation program 309 is activated by a request from the virtual machine performance management client 203, periodic execution of a period specified in advance, and pressing of the data update button 704 or performance analysis button 705 by the user. A performance analysis screen 701 is created and transmitted to the virtual machine performance management client 203. Then, the data collection program 211 is called by pressing the data update button 704 or periodic execution, and the performance drop cause analysis screen creation program 310 is called by pressing the performance drop virtual machine name button 708.

  The data collection program 211 is activated by a call from the virtual machine performance analysis screen creation program 309 or by periodic execution of a predetermined cycle, acquires the latest information from the virtual environment management server 201, and stores the resource usage record storage table 313, the configuration information storage table 314, and the configuration change event storage table 315 information are updated.

  The performance degradation cause analysis screen creation program 310 is activated when the performance degradation virtual machine name button 708 or the configuration display button 806 is pressed, creates a performance degradation cause analysis screen 801, and transmits it to the virtual machine performance management client 203. When the return button 807 is pressed, the virtual machine performance analysis screen creation program 309 is called.

<Embodiment 1: Data collection program>
FIG. 10 is a diagram for explaining the processing flow of the data collection program 211. Hereinafter, each step of FIG. 10 will be described.

  The data collection program 211 checks whether it is regularly executed in a predetermined cycle, and if it is periodically executed, the process proceeds to step S1002. Other than periodic execution (in the first embodiment, a call from the virtual machine performance analysis screen creation program 309) ), The process proceeds to step S1003 (step S1001).

  The data collection program 211 acquires the configuration and status information of all virtual machines, physical machines, and storages managed by the virtual environment management server 201 from the virtual environment management server 201 and stores them in the configuration information storage table 314 ( Step S1002).

  The data collection program 211 acquires from the virtual environment management server 201 the resource usage record information of all virtual machines, physical machines, and storages managed by the virtual environment management server 201 and stores it in the resource usage record storage table 313. (Step S1003). The resource usage record information acquisition period is a period from the previous acquisition date and time until the data collection program 211 is executed.

  The data collection program 211 acquires from the virtual environment management server 201 event information related to the configuration and status change of all virtual machines, physical machines, and storages managed by the virtual environment management server 201, and stores the configuration change event. It stores in the table 315 (step S1004). The event information acquisition period is a period from the previous acquisition date and time to the execution point of the data collection program 211.

<Embodiment 1: Virtual machine performance analysis screen creation program>
FIG. 11 is a diagram for explaining the processing flow of the virtual machine performance analysis screen creation program 309. Hereinafter, each step of FIG. 11 will be described.

  The virtual machine performance analysis screen creation program 309 checks whether it is periodically executed in a predetermined cycle or activated by pressing the data update button 704, and if it is activated by periodically executing or pressing the data update button 704, the process proceeds to step S1102. move on. If the activation is other than pressing of the data update button 704 (in the first embodiment, the performance analysis button 705 is pressed), the process proceeds to step S1103 (step S1101).

  The virtual machine performance analysis screen creation program 309 executes the data collection program 211 (step S1102).

  The virtual machine performance analysis screen creation program 309 acquires the resource use record information of all virtual machines for the analysis period specified in advance from the resource use record storage table 313, and stores it in the variables (step S1103).

  The virtual machine performance analysis screen creation program 309 acquires an input value if there is an input in the detection threshold input field 706, acquires an initial value if there is no input, and stores it in a variable as detection threshold information. Then, the resource usage record analysis program 212 is executed using the virtual machine resource usage information acquired in step S1103 and the acquired detection threshold information as arguments, and the performance statistical information and the performance-decreasing virtual machine name information are returned as return values. Obtain (step S1104). Here, the performance statistical information is information that is statistically processed in order to grasp a list of performance states of a large number of virtual machines. In the first embodiment, the performance average value for each virtual machine and each resource type in a specified period is obtained. This is the aggregate frequency distribution. The degraded performance virtual machine name information is information including the virtual machine name extracted with reference to the specified detection threshold information and the resource type name that is the detection condition.

  The virtual machine performance analysis screen creation program 309 uses the performance statistical information and the performance-reduced virtual machine name information acquired in step S1104 to create a virtual machine performance trend display area 702 and a performance-reduced virtual machine name display area 703, respectively. The virtual machine performance analysis screen 701 is created. Then, the created virtual machine performance analysis screen 701 is transmitted to the virtual machine performance management client 203 and displayed. (Step S1105). The degraded performance virtual machine name information is displayed as a degraded performance virtual machine name button 708. In addition, the virtual machine performance analysis screen creation program 309 waits for a user operation after executing this step.

  When the user presses the data update button 704 or the performance analysis button 705, the virtual machine performance analysis screen creation program 309 acquires the input value of the detection threshold value input field 706, stores it in the variable, and returns to step S1101. With this process, the user can redisplay the virtual machine performance analysis screen 701 at an arbitrary detection threshold. When the performance degradation virtual machine name button 708 is pressed, the virtual machine performance analysis screen creation program 309 calls the performance degradation cause analysis screen creation program 310 with the performance degradation virtual machine name of the pressed button as an argument and ends ( Step S1106).

<Embodiment 1: Performance degradation cause analysis screen creation program>
FIG. 12 is a diagram for explaining the processing flow of the performance degradation cause analysis screen creation program 310. Hereinafter, each step of FIG. 12 will be described.

  The performance degradation cause analysis screen creation program 310 acquires the virtual machine name specified by the argument, and stores it in a variable as the performance degradation virtual machine name (step S1201).

  The performance degradation cause analysis screen creation program 310 obtains the resource usage record information of the performance degraded virtual machine name acquired in step S1201 from the resource use record storage table 313, and creates a performance time series graph (step S1202). Here, the performance value type of the resource use record to be acquired is the same as that acquired in step S1103. Further, the acquisition period of the resource usage record information to be acquired is a period from a certain period before the input normal operation date and time (for example, 5 minutes before) to the current date and time if a value is input in the normal operation date and time input field 804. If no value is input, the period is the same as that acquired in step S1103.

  The performance degradation cause analysis screen creation program 310 acquires an input value if there is an input for each of the normal operation date and time input field 804 and the resource type input field 805, and acquires an initial value if there is no input, and the normal operation date and time Information and resource type information are stored in variables. Then, the system configuration information analysis program 213 is executed using the acquired normal operation date and time information and resource type information, and the performance-reduced virtual machine name acquired in step S1201 as arguments, and the system configuration information (hereinafter, the specified normal operation date and time) Is described as normal operating time system configuration information) (step S1203). Here, the system configuration information is information including a configuration map of a shared resource used by the performance-reduced virtual machine, a power state, and resource usage load information. Note that the resource usage load information is information used to display the resource usage amount display sign 814, and is information in which the level of the resource usage load, such as level 1 to level 3, is expressed in stages according to a predetermined threshold value.

  The performance degradation cause analysis screen creation program 310 executes the system configuration information analysis program 213 using the performance degraded virtual machine name acquired in step S1201, the current date and time, and the resource type information acquired in step S1203 as arguments, and System configuration information (hereinafter referred to as current system configuration information) is acquired (step S1204).

  The performance degradation cause analysis screen creation program 310 draws the virtual machine performance time-series area 802 using the performance time-series graph acquired in step S1202, and further the normal operation point system configuration information acquired in step S1203 and the step S1204. Using the acquired current system configuration information, the shared resource usage status display area 803 is drawn to create a performance degradation cause analysis screen 801. The performance degradation cause analysis screen creation program 310 transmits and displays the created performance degradation cause analysis screen 801 to the virtual machine performance management client 203 (step S1205). The performance degradation cause analysis screen creation program 310 waits for a user operation after executing this step. Note that the display of the shared resource usage status display area 803 highlights the performance-reduced virtual machine acquired in step S1201. For example, in this embodiment, a performance-reduced virtual machine is placed at the top of the system configuration map 812 and displayed in a different color from other virtual machines.

  When the user presses the configuration display button 806, the performance degradation cause analysis screen creation program 310 acquires the input values of the normal operation date / time input field 804 and the resource type input field 805, stores them in variables, and step S1202 Return to. By this processing, the user can redisplay the performance degradation cause analysis screen 801 for the arbitrary normal operation date and time and the resource type. When the return button 807 is pressed, the performance degradation cause analysis screen creation program 310 calls the virtual machine performance monitoring screen creation program 309 and ends (step S1206).

<Embodiment 1: Resource usage record analysis program>
FIG. 13 is a diagram for explaining the processing flow of the resource usage record analysis program 212. Hereinafter, each step of FIG. 13 will be described.

  The resource usage record analysis program 212 acquires virtual machine resource use record information and detection threshold information from the argument, and stores them in variables (step S1301).

  The resource usage record analysis program 212 totals the virtual machine resource usage record information acquired in step S1301, creates virtual machine performance trend total information, and stores it in a variable (step S1302). In the first embodiment, as the virtual machine performance trend total information, the performance average values for each virtual machine and each resource type in a certain period are totaled to create a frequency distribution.

  The resource usage record analysis program 212 graphs the virtual machine performance trend total information acquired in step S1302, and creates performance statistical information (step S1303). In the first embodiment, image data is created by creating a frequency distribution diagram.

  The resource usage record analysis program 212 identifies the name and resource type name of the virtual machine whose aggregate value exceeds the value of the detection threshold information acquired in step S1301 among the virtual machine performance trend aggregate information created in step S1302. The virtual machine name and the resource type name are stored in a variable as performance degradation virtual machine name information (step S1304). In addition, this step may perform the following operation when an empty character which is an initial value is designated in the detection threshold information. From the virtual machine performance trend summary information created in step S1302, the virtual machine name up to the 10th worst performance value is identified for each resource type, and the virtual machine name and resource type name are stored in a variable as performance degradation virtual machine name information. To do.

  The resource usage record analysis program 212 returns the performance statistical information (frequency distribution data and frequency distribution image data) acquired in step S1303 and the performance-reduced virtual machine name information acquired in step S1304, and ends (step S1305). ).

<Embodiment 1: System configuration information analysis program>
FIG. 14 is a diagram for explaining the processing flow of the system configuration information analysis program 213. The process of the system configuration information analysis program 213 is roughly divided into two processes. Steps S1401 to S1404 reproduce the system configuration of the performance-reduced virtual machine at the date and time specified by the argument, and use the shared resources (the physical machine 204 and the storage 206 in the first embodiment) used by the virtual machine. This is the process to identify. Steps S1405 to S1409 acquire configuration maps, power supply statuses and resource usage load statuses of all physical machines / virtual machines that have been operating on the shared resource specified in the processing up to step S1404, and create return values. It is processing to do.

  In the first embodiment, the following measures are taken in order to reduce the influence that the acquisition of data necessary for reproducing the system configuration has on the virtual environment. The configuration information storage table 314 holds snapshot-like information of the entire system configuration at a certain point in time. On the other hand, the configuration change event storage table 315 holds only configuration change events that have occurred in the system since the previous data acquisition. In order to reproduce a system configuration at an arbitrary time point using only the configuration information storage table 314, the configuration information storage table 314 must be updated frequently at short time intervals. Since the statuses of all system components managed by the environment management server 201 are referred to, a high load is applied to the virtual environment management server 201 and the physical machine 204 when collecting data. Therefore, in the first embodiment, the system configuration is reproduced by applying the event information held in the configuration change event storage table 315 to the snapshot of the configuration held in the configuration information storage table 314. According to this method, since the data update frequency of the configuration information storage table 314 can be suppressed, the system configuration at an arbitrary time can be reproduced without imposing a high load on the virtual environment.

  The system configuration information analysis program 213 acquires the performance-reduced virtual machine name, the configuration display date and time, and the resource type information from the arguments and stores them in variables (step S1401). The configuration display date and time acquired in this step is hereinafter referred to as configuration display date and time T1 in this paper.

  The system configuration information analysis program 213 obtains the latest record from the configuration information storage table 314, in which the virtual machine name 3144 matches the performance-reduced virtual machine name acquired in step S1401 and the information collection date 3141 is before the configuration display date T1. This is retrieved and stored in the variable as base configuration information K1. The information collection date 3141 of the acquired base configuration information K1 is stored in a variable as the base date T2 (step S1402). This step is for acquiring the system configuration information stored in the configuration information storage table 314 immediately before the date and time T1 for displaying the system configuration. Since it is sufficient to apply the configuration change event information as a difference to the base configuration information K1 after the system configuration information after T2, the latest system configuration information is first acquired in this step. However, if the system configuration information at the date T1 exists, it can be used as it is, so the following steps are not necessary.

  From the configuration change event storage table 315, the system configuration information analysis program 213 includes the entity name 3152 in which the performance-reduced virtual machine name acquired in step S1401 is included, and the event occurrence date 3151 is from the base date T2 to the configuration display date T1. Are searched for, and these records are stored in a variable as configuration change history information K2 (step S1403).

  The system configuration information analysis program 213 merges the base configuration information K1 acquired in step S1402 and the configuration change history information K2 acquired in step S1403 to generate configuration information K3 of the degraded performance virtual machine, and displays the configuration display date and time. The shared resource name (used physical machine name, used storage name) used by the virtual machine whose performance has deteriorated at time T1 is specified (step S1404).

  The system configuration information analysis program 213 searches the configuration information storage table 314 for the newest record in which the storage name 3142 matches the used storage name specified in step S1404 and the information collection date 3141 is before the configuration display date T1. This is stored in a variable as shared resource base configuration information K4. Further, the physical machine name, virtual machine name, and information collection date / time included in the shared resource base configuration information K4 are stored in variables as the shared physical machine name, shared virtual machine name, and shared resource base date / time T3, respectively (step S1405). This step is for extracting all physical machines and virtual machines that have used the storage 206. That is, even when the storage 206 is shared across physical machines, it is meaningful to extract the usage relationship.

  The system configuration information analysis program 213 adds the shared physical machine name or shared virtual machine name acquired in step S1405 to any one of the entity name 3152, the migration source entity name 3154, and the migration destination entity name 3155 from the configuration change event storage table 315. Are included and the event occurrence date / time 3151 is in the range from the shared resource base date / time T3 to the configuration display date / time T1, and these records are stored in the variable as the shared resource configuration change history information K5 (step S1406). .

  The system configuration information analysis program 213 merges the shared resource base configuration information K4 acquired in step S1405 and the shared resource configuration change history information K5 acquired in step S1406 to generate shared resource configuration information K6, and displays the configuration display date and time. The configuration map and power supply state of the performance-reduced virtual machine, shared physical machine, and shared virtual machine at T1 are specified (step S1407).

  From the resource usage record storage table 313, the system configuration information analysis program 213 obtains resource usage record data in which the resource type 3133 matches the resource type information acquired in step S1401 and the entity name 3152 is included in the shared resource configuration information K6. Search and get. Further, the data is totaled according to the threshold value of the resource usage amount display sign 813 specified in advance, and resource usage load information is generated (step S1408).

  The system configuration information analysis program 213 returns the shared resource configuration information K6 generated in step S1407 and the resource usage load information generated in step S1408, and ends the processing (step S1409).

<Embodiment 1: Summary>
As described above, the virtual environment operation support system 1000 according to the first exemplary embodiment records the history of usage relations between resources (physical devices) shared between physical machines and virtual machines as system configuration information. In the performance degradation cause analysis screen 801, the usage relationship between the virtual machine and the resource at a plurality of times is displayed together with the performance performance value of the virtual machine at each time. As a result, even when resources are shared across physical machines, it is possible to identify the influence of changes in the system configuration on changes in the performance of virtual machines.

<Embodiment 2>
In the first embodiment, the physical machine and the storage have been described as examples of the shared resource. As another example of shared resources, for example, the same management may be performed for network devices. The network device here is a device that exists on the network and is commonly used by each physical machine, such as a router and a LAN switch.

  In this case, for example, the network device management configuration information storage table in which the storage name 3142 portion of the configuration information storage table in FIG. 5 is changed to the network device name is added, and the data collection program 211 stores the resource usage record of the network device. Information, system configuration information, and system configuration change event information are also acquired. In addition, in the virtual machine performance analysis screen 701, an input field for selecting whether to display a storage configuration or a network configuration as a shared resource is added, and when the storage configuration is specified in the input field, When the system configuration information analysis program 213 described in FIG. 12 is executed and the network configuration is specified in the input field, the system configuration information is referred to by referring to the network device management configuration information storage table instead of the configuration information storage table. Assume that the analysis program 213 is executed.

<Embodiment 3>
In the performance degradation cause analysis screen creation program 310 described with reference to FIG. 12, the date and time input in the normal operation date and time input field 804 is compared with the shared resource usage status at the current date and time. However, the comparison of the shared resource usage status may be performed at an arbitrary date and time input by the user.

  In this case, for example, an abnormal operation date / time input field and an abnormal operation date / time sign having the same functions as the normal operation date / time input field 804 and the normal operation date / time sign 809 are added to the performance degradation cause analysis screen 801. In step S1204 of the performance deterioration cause analysis screen creation program 310, processing is performed using the date and time input in the abnormal operation date and time input field instead of the current date and time, and system configuration information (abnormal operation point system configuration information) is executed. Shall be obtained. In step S1205, processing is performed using the abnormal operation time system configuration information instead of the current system configuration information. In step S1206, in addition to the input values of the normal operation date / time input field 804 and the resource type input field 805, the value of the added abnormal operation date / time input field is added to perform the process.

<Embodiment 4>
A configuration difference log display button for presenting the difference in the system configuration information displayed in the shared resource usage status display area 803 as a log to the user may be added to the performance degradation cause analysis screen 801.

  In this case, for example, when the configuration difference log display button is pressed in step S1206, the performance deterioration cause analysis screen creation program 310 determines that the event occurrence date / time 3151 is specified in step S1203 from the configuration change event storage table 315. It is assumed that records in the range from the operation date / time to the date / time specified in step S1204 are searched, and a screen for presenting these records to the user is created and displayed.

201 Virtual environment management server 202 Virtual machine performance management server 203 Virtual machine performance management client 204 Physical machine 206 Storage 209 Virtual machine 211 Data collection program 309 Virtual machine performance analysis screen creation program 310 Performance degradation cause analysis screen creation program 212 Resource usage record analysis Program 213 System configuration information analysis program 313 Resource usage record storage table 314 Configuration information storage table 315 Configuration change event storage table 1000 Virtual environment operation support system

Claims (14)

A system that supports the operation of a virtual machine built on a physical machine,
Stores system configuration information describing the history of usage relationships between physical devices shared by a plurality of the physical machines and the virtual machines, and resource usage record information recording the actual use of the physical devices by the virtual machines A table part to be
An arithmetic processing unit using information acquired from the table unit;
With
The arithmetic processing unit includes:
Accepting a designation for the virtual machine to output the usage status of the physical device and a designation for two dates and times to output performance performance values of the designated virtual machine;
Based on the description of the system configuration information and the resource usage record information, the usage status of the physical device used by the specified virtual machine at each of the two specified dates and times is specified, and each specified date and time is specified. The virtual environment operation support system characterized in that the usage status and performance performance value of the virtual machine at each date and time are output. The arithmetic processing unit includes:
Based on the resource usage record information, calculate and output statistical information about the performance record of the virtual machine for each physical device type,
Accepting a specification for a performance threshold for identifying the virtual machine in a degraded state,
Based on the specified performance threshold value and the statistical information, the virtual machine in a state where the performance has deteriorated is identified, and the identifier of the virtual machine is ranked according to the seriousness of the performance deterioration and output. The virtual environment operation support system according to claim 1, wherein: The arithmetic processing unit includes:
As the statistical information, the performance results for each type of the physical device in a predetermined period are averaged over the predetermined period,
The average value is used as a data range, and a frequency distribution with the number of virtual machines that fall within each data range as a frequency is output,
The virtual environment operation support system according to claim 2, wherein any value in the data range on the frequency distribution is accepted as designation for the performance threshold. The arithmetic processing unit includes:
Accepting a specification for the performance threshold for each type of the physical device,
The virtual environment operation support system according to claim 2 or 3, wherein an identifier of the virtual machine exceeding the performance threshold is output for at least one of the physical device types. The arithmetic processing unit includes:
The degree of performance degradation of the virtual machine is ranked based on the performance performance for each type of the physical device, and the identifiers of the virtual machines up to a predetermined order counted from the higher degree of performance degradation are output. The virtual environment operation support system according to claim 4. The arithmetic processing unit includes:
6. The ranking is performed so that the degree of performance degradation increases as the number of types of physical devices exceeding the performance threshold increases and the degree of performance degradation increases. The virtual environment operation support system described. The arithmetic processing unit includes:
When a specification for the date and time is received,
2. A time series graph of performance performance values of the virtual machine in a period including the designated date and time, and information indicating a position of the designated date and time on the time series graph are output. The virtual environment operation support system according to any one of items 1 to 6. The arithmetic processing unit includes:
The specification for instructing the two date and time is a current date and time and a specification indicating a date and time prior to the current date and time, and specifies and outputs the current status and the usage status at each of the specified date and time. Item 8. The virtual environment operation support system according to any one of Items 1 to 7. The arithmetic processing unit includes:
The specification for the two dates and times is received by specifying two dates and times prior to the current date and time, and the usage status at each of the two specified dates and times is specified and output. 8. The virtual environment operation support system according to any one of items 1 to 7. The arithmetic processing unit includes:
About the physical device that was used at each of the two dates and times when the virtual machine was specified, and all the other virtual machines that were using the physical device,
A system configuration map showing the connection relationship between each virtual machine and the physical device, status information showing the power status of each virtual machine, and usage information relatively showing the magnitude of the actual usage value of the physical device The virtual environment operation support system according to any one of claims 1 to 9, wherein the virtual environment operation support system is calculated and output as the usage status. The arithmetic processing unit includes:
About the physical device that was used at each of the two dates and times when the virtual machine was specified, and all the other virtual machines that were using the physical device,
Calculating event information describing a change in the usage relationship that occurred between the two dates and times and a change in the power state of each virtual machine, and outputting the event information instead of the usage status or together with the usage status The virtual environment operation support system according to any one of claims 1 to 10. The table portion is
Storing configuration change event information describing the change history of the usage relationship;
The arithmetic processing unit includes:
From the system configuration information, the usage relationship at a time point closest to the date and time before the older one of the specified date and time is acquired,
The usage status at the two specified dates and times is calculated by sequentially applying the configuration change event information in the period between the two specified dates and times to the acquired usage relationship. The virtual environment operation support system according to any one of claims 1 to 11. The virtual environment operation support system according to any one of claims 1 to 12, wherein the physical device is a storage device shared by a plurality of the physical machines. The virtual environment operation support system according to any one of claims 1 to 12, wherein the physical device is a network device shared by a plurality of the physical machines.

Images