JPH06110715A - Dynamic allocation method of computer resources in virtual computer system - Google Patents

Abstract

(57) [Summary] [Objective] To provide a virtual computer system capable of dynamically changing the allocation of a main storage area to an OS. [Structure] Taking the occurrence of a specific event as an opportunity, the area of the high-order address in the guest main memory area of the virtual machine VM-2 is made offline (not connected) (step 701).
0). Next, the virtual machine monitor 410 moves the guest main storage area of the VM-2 (step 7020) to make the area of the high-order address adjacent to the guest main storage area of the virtual machine VM-1 free. Then, the area of the high-order address adjacent to the guest main memory area of VM-1 is brought online (connected state) and connected to VM-1 (step 7040). In this way, the allocation of the main storage area to the OS is dynamically changed.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a computer system, and more particularly to a computer system in which a plurality of OSs run on one computer, which is suitable for dynamically changing the allocation of computer resources to the OS. Regarding computer systems.

[0002]

2. Description of the Related Art As a computer system capable of running a plurality of operating systems (OS) on one computer, a virtual computer system (VMS:
Virtual Machine System). The virtual computer system generates a plurality of virtual computers (VMs) that are logical computers on a single computer, and enables each OS to run a corresponding OS. This virtual computer system can operate multiple OSs with different purposes on a single computer, perform system generation of a new version of OS while operating an old version of OS, or run on a single computer. It is mainly used to execute OS tests in parallel. In a virtual machine system, a virtual machine monitor (VMM: Virtual Machine Monitor) that is a control program for the virtual machine system.
r) distributes computer resources to each virtual machine (VM),
Performs scheduling and dispatch of virtual machines (VMs). Further, the virtual machine monitor (VMM) performs a simulation process of an instruction issued by the OS on the virtual machine that cannot be directly executed by the computer, that is, an instruction requiring software intervention.

A pageable VM is provided to the virtual machine depending on how the main storage area (guest main storage area) of the virtual machine is given.
And there is a page fixed VM. Pageable VM is Japanese Patent Laid-Open No.
As described in JP-A-7-212680, a guest main storage area is made resident in a virtual space (host virtual space) created by a virtual machine monitor (VMM). Therefore, the guest main storage area of the pageable VM is the virtual machine monitor (VM
M) is the target of paging. By this paging, the size of the guest main memory area of the pageable VM is
The size can be made larger than the size of the paging area of the virtual machine monitor (VMM) that is the mapping destination of the host virtual space.

On the other hand, the page fixing VM is disclosed in JP-A-60-2.
As described in Japanese Patent No. 4735, the guest main memory area is made resident in a continuous area on the main memory device. Therefore, the guest main storage area of the page-fixed VM is the virtual machine monitor (VM
It is not subject to paging by M). Therefore, the fixed page VM requires less overhead for address conversion than the pageable VM. However, page fixing V
As the size of the guest main memory area of M, only the size of the continuous area of the associated main memory device can be set.

Further, Japanese Patent Laid-Open No. 3-211629 discloses that
A virtual computer that temporarily suspends access to the guest main storage area by an asynchronous processing device such as a channel that performs processing asynchronously with the central processing unit, and makes it possible to move the guest main storage area of the page fixed VM on the main storage device. The system is disclosed. According to this virtual computer system, fragmentation of the storage area of the main storage device can be prevented by moving the guest main storage area.

Further, Japanese Laid-Open Patent Publication No. 2-201655 discloses a virtual computer allocation system in which a plurality of central processing units constituting one computer are assigned as the central processing units of a virtual computer.

[0007]

The assignment of the guest main memory area in the virtual computer system has the following problems.

(1) The first problem relates to the total balance of the sizes of the guest main memory areas.

In the operation of the virtual computer system, load fluctuations of the virtual computer often occur regularly. For example,
When the two virtual machines are operated all day long, the first virtual machine operates a TSS (Time Sharing System) and thus has a heavy daytime load.
In many cases, the night load is high because batch jobs are executed in free time. In such a case, giving a larger main storage area to the virtual machine with the higher load can reduce the paging process and the like, and improve the overall throughput.

Further, as shown in FIG. 1, the first computer 50
0 to the virtual computer VM-1 on the second computer 600
Is in the hot standby state, the first computer 50
While 0 is operating normally, a relatively small amount of guest main storage area 420-1 is given to the virtual machine VM-1,
It is preferable to enlarge the guest main storage area 420-2 of the virtual machine VM-2 that operates batch jobs and TSS. This means that in the virtual machine VM-1, the first computer 5
This is because only the processing for the case where a failure occurs in 00 is performed. When the system switching device 700 switches the processing of the first computer 500 to the virtual computer VM-1 due to the failure of the first computer 500, the size of the guest main storage area 420-1 of the virtual computer VM-1 is increased. It is desirable to be able to withstand high loads.

However, conventionally, in order to change the size of the guest main storage area of a virtual machine, the arrangement of the main storage areas of other virtual machines must also be changed, so that all virtual machines must be stopped. There wasn't.

(2) The second problem relates to the movement of the guest main storage area on the main storage device. A channel program (guest channel program) requested by the OS on the virtual machine to be executed by an input / output instruction is described as an absolute address (guest absolute address) in the guest main storage area. Therefore, the virtual machine monitor (V
In the simulation by MM), a channel program (host channel program) in which a data address is converted into an absolute address (host absolute address) in the main memory is created, and then this host channel program is executed. Therefore, all the data addresses of the host channel programs that started to be executed before the movement of the guest main memory area indicate the guest main memory area before the movement. Therefore, when the guest main storage area is moved by the method disclosed in Japanese Patent Laid-Open No. 3-211629 and the execution of the host channel program is restarted by the virtual machine monitor, only the data in the area before the movement is updated and after the movement. The data in the guest main memory area is not updated. Therefore, data inconsistency occurs.

(3) The third problem relates to the change from the page fixed VM to the pageable VM. When a main storage area is required due to the creation of a new virtual machine, page fixed VM
As a pageable VM, it is desirable to create an empty area. At this time, since the guest main storage area is changed in size, moved, etc., the same problem as described above occurs.

(4) The fourth problem relates to the change from the pageable VM to the page fixed VM. When there is a sufficient free area on the main storage device, it is desirable to set the pageable VM to a page fixed VM in order to improve the performance. Also in this case, the size and movement of the guest main storage area are usually involved, and therefore the same problem as described above occurs.

(5) The fifth problem relates to the allocation of the central processing unit to the virtual machine. When the load fluctuation of the virtual machine occurs regularly or when the system is switched to the virtual machine in the hot standby state, the central processing unit should be given priority to the virtual machine with the high load. However, as disclosed in Japanese Patent Laid-Open No. 2-201655, conventionally, allocation of a central processing unit to a virtual computer has been fixedly determined when the virtual computer was created.

An object of the present invention is to provide a method of dynamically allocating computer resources in a virtual computer system, which allows reallocation of computer resources according to load fluctuations of each virtual computer.

Another object of the present invention is to provide a virtual computer system capable of dynamically changing the size of the guest main memory area.

Still another object of the present invention is to provide a virtual computer system in which a guest main memory area can be moved on the main memory device without stopping access to the guest main memory area by an asynchronous processing device such as a channel. It is to provide a method for dynamically allocating computer resources.

Another object of the present invention is to provide a method for dynamically allocating computer resources in a virtual computer system, which can dynamically change the allocation of a central processing unit to a virtual computer.

[0020]

A method of allocating computer resources in a virtual computer system according to the present invention comprises a central processing unit, a main storage unit, and an asynchronous processing unit for the central processing unit in response to an instruction from the central processing unit. In a virtual computer system in which a plurality of operating systems (OS) run under the control of a virtual computer monitor on a computer including an asynchronous processing device that performs processing, among the plurality of OSs,
The first fixed area of the main storage device is allocated to the first OS, and the second fixed area adjacent to the higher address side of the first fixed area is allocated to the second OS. Step and when a specific event occurs,
Reducing the size of the second fixed area allocated to the second OS, and moving the second fixed area having the reduced size to the higher address side on the main storage device, The method includes a step of generating a free area on the higher address side of the fixed area No. 1 and a step of expanding the size of the main storage area of the first OS.

In the step of generating the free area, preferably, the main storage area of the second OS is a virtual space generated by the virtual machine monitor, and the virtual space is the reduced second fixed area. The step of mapping, the step of changing the mapping destination of the virtual space to the area of the movement destination of the main storage device, and the step of changing the area of the movement destination to the second area.
And assigning it as a fixed area of the OS.

During the movement of the main storage area of the second OS, the virtual computer monitor simulates a use start instruction of the asynchronous processor issued by the second OS, and A step of analyzing which area of the main memory area of the second OS is accessed by the asynchronous processing device; and a step of moving the area to be accessed after the access by the asynchronous processing device is completed. It is desirable to have.

[0023]

[Action]

(1) When creating each virtual machine, the computer resources of this virtual machine before and after the occurrence of a specific event are defined in the configuration table. Then, when a specific event occurs, the size of the main storage area (guest main storage area) of the second OS (virtual computer) is reduced according to the definition of the configuration table, and the guest main storage area of each virtual computer is stored in the main storage area. By moving on the device, the area with the higher address adjacent to the guest main memory area of the first virtual machine is made a free area. After that, the size of the guest main memory area of the first virtual computer is expanded according to the definition of the configuration table.

Further, at this time, the allocation of the central processing unit to the virtual machine can be changed according to the definition of the configuration table. For example, from the state where three virtual computers shared two central processing units, the first and second virtual computers shared the first central processing unit, and the first virtual computer shared the second central processing unit. Transition to a state where the central processing unit is occupied.

Here, the specific event means that an event that a specific time has come or a failure occurs in a second computer different from the computer on which the virtual computer runs, and the processing of this second computer is performed by the first Shows an event that the OS on the virtual computer of the above took over.

As described above, it is possible to provide a virtual computer system capable of dynamically changing the size of the guest main storage area.

Further, it is possible to provide a virtual computer system capable of dynamically changing the allocation of the central processing unit to the virtual computer.

(2) When the guest main memory area of the page-fixed VM is moved on the main memory device, the virtual machine monitor (VMM) creates the correspondence between the guest main memory area and the continuous area on the main memory device. Defined in the virtual space (host virtual space), the mapping destination of the host virtual space is changed to the migration destination area, and the continuous area on the main storage device after the migration is allocated as the guest main storage area of the OS. Also, while the guest main memory area is being moved, the virtual machine monitor (VMM) simulates an instruction to start using the asynchronous processing device by the OS, and the access target area of the guest main memory area by the asynchronous processing device is after the access is completed. ,Moving. At this time, the use of the pageable VM allows the main storage area corresponding to the host virtual space to be discontinuous in page units while the guest main storage area is being moved. In this way, the areas are sequentially moved from the area outside the access target of the guest main storage area by the asynchronous processing device. For this reason,
After the area is moved, the asynchronous processing device does not accidentally access the area before the movement.

As described above, the guest main storage area can be moved in the main storage device without stopping the access to the guest main storage area by the asynchronous processing device such as the channel.

(3) The correspondence between the guest main memory area of the page-fixed VM and the continuous area on the main memory device in which it is resident is defined in the host virtual space, and the definition information of the host virtual space is used for address translation to the OS. Run. Further, the area size of the host virtual space mapping destination is paged out to make the area size of the host virtual space mapping destination smaller than the guest main memory area size. In this way, the page fixed VM can be dynamically changed to the pageable VM.

(4) If the host virtual space mapping destination area that holds the correspondence between the guest main memory area of the pageable VM and the main memory device is paged out, this area is paged on the main memory device. In. Next, as a mapping destination of the host virtual space, a continuous area of the main storage device is associated. Then, the OS is run using the arrangement information in the main memory of this continuous area. In this way, the pageable VM can be dynamically changed to the page fixed VM.

[0032]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A first embodiment relating to a guest main memory area moving method and a second embodiment relating to a computer resource allocation changing method according to the present invention will be described below in detail with reference to the drawings.

I. Method of Moving Guest Main Storage Area FIG. 2 shows a configuration diagram of the virtual machine system (VMS) of this embodiment. In FIG. 2, 100-1 and 100-
2 is a central processing unit, 200 is a central processing unit 100-k
(K = 1, 2) is an asynchronous processing device that performs processing asynchronously, and 300 is a storage control device that controls reference and update to the main storage device 400.

The asynchronous processor 200 is a channel when the computer is an ordinary general-purpose large computer. The channel transfers data between the input / output device and the main memory 400 asynchronously with the operation of the central processing unit 100-k (k = 1, 2). When the computer is a super computer, the asynchronous processing device 200 is a vector processor. The vector processor is a central processing unit 100-
Asynchronously with the operation of k (k = 1, 2), the main storage device 400
The above instruction sequence (so-called vector instruction sequence) is executed.
In this embodiment, three page fixed VMs are running as the initial state, and the storage area of the main storage device 400 excluding the area of the virtual machine monitor (VMM) 410 is the virtual machine V.
Guest main memory area 420-1 of M-1, virtual machine VM
-2 guest main storage area 420-2, virtual machine VM-
It is assumed that the guest main storage area 420-3 of No. 3 is allocated to each virtual computer.

Next, the virtual computer monitor 41 will be described with reference to FIG.
The control block 430-i (i = 1, 2, 3) created by 0 to control the running of each virtual computer will be described. The control block 430-i is for each virtual machine VM-
i (i = 1, 2, 3) exists and the virtual computer V
The OS on the M-i is the central processing unit 100-k (k = 1,
In 2), various control information during running is given.

In FIG. 3, the direct execution bit 431 is 1
When, the start instruction of the asynchronous processing device 200 issued by the OS on the virtual machine is directly executed, and when 0, the execution of the instruction is suppressed and executed via the virtual machine monitor 410. When the page-fixed VM identification bit 432 is 1, it indicates that the virtual machine is a page-fixed VM, and when it is 0, it indicates that it is a pageable VM. When the page fixed VM identification bit 432 is 1, the origin address field 433 and the end address field 43
Reference numeral 4 holds the starting point address and the ending point address in the main memory 400 of the guest main memory area. When the page fixed VM identifier 435 is 0, the host first control register 435 holds the start address of the host segment table (host ST) that defines the host virtual space in which the guest main storage area of this virtual computer resides. To do.

Next, address conversion in the virtual machine system (VMS) will be described with reference to FIG.

As for the logical address (guest logical address) 4000 in the virtual space (guest virtual space) generated by the OS on the virtual machine VM-2, the OS is the guest main storage area 42.
0-2, the guest segment table (guest ST) and guest page table (guest PT) created on the result of address conversion 4100, guest absolute address 4
Converted to 200. Then, when the virtual machine is the page-fixed VM, that is, the control block 43 of this virtual machine
When the page fixed VM identification bit 432 of 0 is 1, the conversion result by the page fixed VM address conversion circuit 4300 is validated by the selector 4500 and becomes the host absolute address 4600. On the other hand, virtual machines can page
When M, that is, the control block 430 of this virtual computer
When the page fixed VM identification bit 432 of 0 is 0, the conversion result by the page conversion possible VM address conversion circuit 4400 is validated by the selector 4500 and becomes the host absolute address 4600.

Here, the conversion from the guest absolute address 4200 to the host absolute address 4600 by the page fixed VM address conversion circuit 4300 is described in the above-mentioned JP-A-60-.
As disclosed in Japanese Patent No. 24735, this is performed by adding the value of the origin address field 433 to the guest absolute address 4200 by the adder 4310. However, when the value of the addition result exceeds the value of the end point address field 434, the comparator 4320 causes the interrupt signal 4330 to report the program interrupt to the OS.
Generate.

On the other hand, the conversion from the guest absolute address 4200 to the host absolute address 4600 by the pageable VM address conversion circuit 4400 is described in JP-A-57-2.
As disclosed in Japanese Patent No. 12680, it is performed as follows. First, the guest absolute address 4200 is a host virtual address when viewed from the virtual machine monitor 410. Therefore, the value of the host first control register 435 and the value of the segment field (S) of the host virtual address 4410 are added by the adder 4420 to obtain the entry address of the host segment table (host ST) 4430. Page field of this entry and page field of host virtual address 4410 (P)
Is added by the adder 4440 to obtain the entry address of the host page table (host PT) 4450. Further, by combining the page address of this entry and the displacement (D) of the host virtual address 4410, the host absolute address 4600 is obtained.

Next, a method for moving the main storage area of the virtual machine on the main storage device 400 will be described with reference to FIGS. 5 and 6. FIG. 5 is a flowchart of the guest main memory area moving method, and FIG. 6 is an explanatory diagram of the guest main memory area moving method.

It is assumed that the virtual machine VM-3 has finished running in the initial state shown in FIG. 6A, and as shown in FIG. 6B, a partial area of the main storage device 400 has become empty. . Therefore, the guest main storage area 420-2 of the virtual machine VM-2 is set to the main storage device 4 by the method described below.
Move on 00.

First, the virtual machine monitor 410 operates as the asynchronous processor 200 issued by the OS on the virtual machine VM-2.
The direct execution bit 431 in the control block 430-2 of the virtual machine VM-2 is set to 0 (step 5000) in order to simulate the boot start instruction of the virtual machine VM-2. As a result, when the OS on the virtual machine VM-2 issues a boot start command for the asynchronous processing device 200 while the guest main memory area 420-2 is moving, the execution of the command is suppressed and the virtual machine monitor 410 is allocated. Be issued.

Next, the virtual machine monitor 410 sends the virtual machine VM-2, which is the page-fixed VM, to the virtual machine monitor shown in FIG.
As shown in (c), a host virtual space 450 is created,
Change VM to pageable VM (step 501)
0). A host segment table (host ST: Host Segment Table) and a host page table (host PT: Host Page Ta) that define this host virtual space 450.
ble) sets the table value such that the value obtained by adding the starting address of the paging area shown in FIG. 6C to the host virtual address becomes the host absolute address of the guest main storage area of the virtual machine VM-2, Virtual machine monitor 410
Create in the area inside. That is, the virtual computer monitor 410 is configured so that the j-th page of the host virtual space 450 is associated with the j-th page of the paging area shown in FIG. 6C.
Create host ST and host PT.

Next, the virtual machine monitor 410 determines whether the size of the mapping destination area needs to be reduced (step 5020). This determination is made based on whether or not there is a command to input the command of the user of the virtual machine system (VMS) to reduce the area size of the mapping destination. If specified, the area size of the mapping destination is changed to the specified size, and the remaining fields are paged out. As a result, the size of the mapping destination area of the host virtual space 450 is made smaller than the size of the guest main storage area (step 5030).
However, the guest main storage area to be accessed by the asynchronous processing device 200 is excluded from the page-out target.

Next, it is determined whether the OS to be run is using the asynchronous processor 200 (5035). If it is not in use, the mapping destination of all the areas of the host virtual space of the OS is changed to the movement destination area (5040), and the process proceeds to step 5100. To change the destination area, the contents of the area before the movement are copied to the area after the movement, and the host ST
By changing the host PT, the area on the host virtual space 450 that was mapped to the area before the movement is mapped to the area after the movement. Step 5
If it is being used in 035, the page fixed VM identification bit 43 of the control block 430-2 of the virtual machine VM-2 of FIG.
2 set to 0, host first control register field 4
By setting the address of the host segment table (host ST) in 35, the definition information of the host virtual space 450 is specified and the OS on the virtual machine VM-2 is dispatched (step 5050). After that, the running of the OS (step 5060) is started. While this OS is running, the page fixed VM identification bit 432 in the control block 430-2 of the virtual machine VM-2 is 0, so the conversion from the guest absolute address to the host absolute address is performed in the control block of the virtual machine VM-2. This is performed according to the host segment table (host ST) designated in the host first control register field 435 of 430-2 and the host page table (host PT) designated by this host ST.

After the OS starts running, when the OS on the virtual machine VM-2 issues a start start command for the asynchronous processor 200 (step 5070), the control block 43 of the VM-2.
Since the direct execution bit 431 of 0-2 is 0, this instruction is determined by the virtual machine monitor 410. As a result, the virtual machine monitor 410 displays the asynchronous processing device 200.
The start-up start instruction is simulated (step 5080). In this simulation, the address of the data transfer target is converted according to the current mapping state of the virtual space, and the activation start instruction is issued again to the asynchronous processing device 200. If the startup is successful (508
3) proceeds to step 5085, and if activation fails due to the reason that the asynchronous processing device is still in use, step 505
Return to 0.

Thereafter, as shown in FIG. 6D, the mapping destination of the host virtual space 450, which is not related to the data transfer of the asynchronous processing device 200, is changed to the migration destination area (step 5085). The area in the guest main storage area 420-2 related to the data transfer of the asynchronous processing device 200 is specified by the start start command of the asynchronous processing device 200.
Therefore, the mapping destination other than the corresponding page is gradually changed to the destination area.

After this simulation, the virtual machine monitor 410 displays the guest main storage area 4 of the virtual machine VM-2.
It is determined whether the movement of 20-2 is completed (step 50
90). This determination can be made based on whether or not the mapping destinations of the host virtual space 450 have all become the destination regions. Then, as shown in FIG. 6E, when the movement is completed, the origin address field 433 in the control block 430-2 is
By setting the starting point address and the ending point address of the moved guest main memory area 420-2 in the and end point address fields 434, the guest main memory device 420-2
The address translation information of (step 510)
0). Next, as shown in FIG. 6F, the virtual machine VM
-2 is set to 1 in the page fixed VM identification bit 432 in the control block 430-2 (step 5110). Also, the OS on the virtual machine VM-2
Control block 430 of the virtual machine VM-2 in order to directly execute the start instruction of the asynchronous processing device 200 by
The direct execution bit 431 in -2 is set to 1 (step 5120).

Then, the OS is dispatched by designating the control block 430-2 of the virtual machine VM-2 of FIG. 3 (step 5130). After that, the running of the OS (step 5140) is started. While this OS is running, since the page fixed VM identification bit 432 in the control block 430-2 of the virtual machine VM-2 is 1 in the conversion from the guest absolute address to the host absolute address, the origin address is set to the guest absolute address. This is done by adding the value of field 433. Further, the upper limit of the address is checked by comparing the value of the end point address field 434 with the host absolute address.

After the start of running, the OS on the virtual machine VM-2
When the start instruction for the asynchronous processor 200 is issued (step 5150), the VM-2 control block 430 is executed.
Since the direct execution bit 431 of -2 is 1, this instruction is
It is directly executed by the instruction processing device 100 without being determined by the virtual machine monitor 410 (step 5).
160).

As described above, according to the present embodiment, the guest main memory is not stopped without stopping access to the guest main memory area by the asynchronous processor 200 such as a channel which performs processing asynchronously with the central processing unit. It is possible to provide a virtual computer system in which an area can be moved on the main storage device. Also,
A page-fixed VM can be dynamically changed to a pageable VM or a pageable VM can be changed to a page-fixed VM.

II. Method of Changing Allocation of Computer Resources A method of changing the allocation of computer resources will be described with reference to FIGS. 7, 8 and 9. FIG. 7 is a flow chart of a method of changing the allocation of computer resources, FIG. 8 is an explanatory view of a method of changing the size of the guest main storage area, and FIG. 9 is an explanatory view of the configuration table 9000. Configuration table 9000
Indicates computer resources to be allocated to each virtual computer before and after the event occurs. For example, the configuration table 9000 of FIG. 9 has a virtual machine VM having a guest main storage area 420-1 of 32 megabytes in the initial state (state before event occurrence) shown in FIG. 8A.
-1 and guest main memory area 4 consisting of 128 megabytes
The three virtual machines, the virtual machine VM-2 having 20-2 and the virtual machine VM-3 having the guest main storage area 420-1 consisting of 32 megabytes, are the central processing unit 10.
It means that the vehicle is running with 0-1 and 100-2 in common.

After that, when a specific event occurs, the virtual machine monitor 410 requests the OS on the virtual machine VM-2 to reduce the guest main storage area 420-2 to 32 megabytes specified in the configuration table ( Step 7
000). Here, the specific event is an event that a specific time has come, or the first computer 50 shown in FIG.
0 occurs and the system switching device 700 causes the processing of the first computer 500 to be performed by the second computer 60.
The event indicates that the OS on the virtual machine VM-1 of 0 has taken over. As a result, the OS on the virtual machine VM-2
Is the guest main storage area 420, as shown in FIG.
-2, the area of the high-order address in -2 is made offline (not connected) (step 7010) and disconnected from the virtual machine VM-2.

Next, the virtual machine monitor 410 moves the guest main storage area 420-2 of the virtual machine VM-2 as in the case of the first embodiment (step 702).
0). As a result, as shown in FIG. 8C, the area of the high-order address adjacent to the guest main storage area 420-1 of the virtual machine VM-1 can be made a free area.

Further, the virtual machine monitor 410 requests the OS on the virtual machine VM-1 to expand the guest main storage area 420-1 to 128 megabytes specified in the configuration table (step 7030). As a result, as shown in (d) of FIG. 8, the OS on the virtual machine VM-1 is
A high-order address area adjacent to the guest main memory area 420-1 is brought online (connected state) (step 704).
0), connect to the virtual machine VM-1.

Further, the virtual machine monitor 410, as shown in the configuration table 9000, has one central processing unit 1.
00-1 is dedicated to the virtual computer VM-1, and the central processing unit 100-2 is used as the virtual computer VM-2 and the virtual computer VM-.
The OS is run on the virtual machine by being shared by the third computer (step 7050).

As described above, according to this embodiment, it is possible to provide a virtual computer system capable of dynamically changing the size of the guest main memory area. Further, it is possible to provide a virtual computer system capable of dynamically changing the allocation of the central processing unit to the virtual computer.

[0059]

According to the present invention, the guest main storage area of each virtual machine is moved on the main storage device, and the area with the higher address adjacent to the guest main storage area of a specific virtual machine is made an empty area. As a result, it is possible to provide a virtual computer system capable of dynamically changing the size of the guest main memory area.

Further, the guest main storage area such as a channel is sequentially moved from an area not to be accessed in the guest main storage area, so that the guest main storage area is not stopped by the asynchronous processing apparatus without stopping access to the guest main storage area. It is possible to provide a virtual computer system in which the main storage area can be moved on the main storage device.

Further, by defining the correspondence between the guest main memory area of the page-fixed VM and the continuous area on the main memory in which it resides in the host virtual space, the page-fixed VM is defined.
It is possible to provide a virtual computer system that can change a page into a pageable VM.

By associating a continuous area of the main storage area as a mapping destination of the host virtual space, it is possible to provide a virtual machine system capable of changing a pageable VM to a page fixed VM.

Furthermore, by allocating the computer resources described in the configuration table to each virtual computer upon the occurrence of an event, it is possible to provide a virtual computer system capable of dynamically changing the allocation of the central processing unit to the virtual computer. .

[Brief description of drawings]

FIG. 1 is an explanatory diagram of a hot standby system for illustrating an application target of the present invention.

FIG. 2 is an explanatory diagram of a virtual computer system to which the present invention is applied.

FIG. 3 is an explanatory diagram of a control block of the virtual computer of FIG.

4 is an explanatory diagram of address conversion in the virtual computer system of FIG.

FIG. 5 is a flowchart of a guest main memory area migration method according to the first embodiment of this invention.

FIG. 6 is an explanatory diagram of a guest main memory area migration method according to the first embodiment of this invention.

FIG. 7 is a flow chart of a method of changing the allocation of computer resources for a virtual computer according to the second embodiment of this invention.

FIG. 8 is an explanatory diagram of a method of changing the size of the guest main memory area according to the second embodiment of this invention.

FIG. 9 is an explanatory diagram of a configuration table according to the second embodiment of this invention.

[Explanation of symbols]

100: central processing unit, 200: asynchronous processing unit, 30
0: Storage control device, 400: Main storage device, 410: Virtual machine monitor, 420-i: Guest main storage area of virtual machine VM-i, 9000: Configuration table

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Masaru Yamamoto 5030 Totsuka-cho, Totsuka-ku, Yokohama, Kanagawa Prefecture Software Development Division, Hitachi, Ltd. (72) Naoko Ikegaya 1099 Ozenji, Aso-ku, Kawasaki, Kanagawa Hitachi Systems Development Laboratory

Claims (9)

[Claims] 1. A plurality of operating systems on a computer including a central processing unit, a main storage unit, and an asynchronous processing unit that performs processing asynchronously with the central processing unit in response to an instruction from the central processing unit. OS) running under the control of a virtual machine monitor, the first fixed area of the main storage device is allocated to the first OS among the plurality of OSs, and the second fixed OS is allocated to the first OS. O
A step of allocating a second fixed area adjacent to the higher address side of the first fixed area with respect to S, and a step of allocating the second fixed area allocated to the second OS when a specific event occurs. A step of reducing the size, and a step of generating an empty area on the higher address side of the first fixed area by moving the reduced second fixed area to a higher address side on the main storage device. And a step of increasing the size of the main storage area of the first OS, the method of allocating computer resources in a virtual computer system. 2. The method according to claim 1, wherein the step of defining the size of the main storage area of each OS before and after the occurrence of the specific event; And a step of changing the size of the main storage area of the OS, the method for dynamically allocating computer resources in a virtual computer system. 3. The method according to claim 1, further comprising a plurality of central processing units, defining a central processing unit to be assigned to each OS before and after a specific event occurs, and the specific event occurs. And a step of allocating a central processing unit to the plurality of OSs according to the above definition, the method for dynamically allocating computer resources in a virtual computer system. 4. The method for dynamically allocating computer resources in a virtual computer system according to claim 1, 2 or 3, wherein the specific event is an event that a specific time has come. 5. The method according to claim 1, 2 or 3, wherein the specific event causes a failure in a second computer different from the computer, and the first OS takes over the processing of the second computer. A method for dynamically allocating computer resources in a virtual computer system, characterized in that it is an event. 6. The step of generating the free space in claim 1 is characterized in that the main storage area of the second OS is a virtual space generated by the virtual machine monitor, and the virtual space is the reduced second space. The method includes the steps of mapping to a fixed area, changing the mapping destination of the virtual space to a movement destination area of the main storage device, and allocating the movement destination area as a fixed area of the second OS. A method for dynamically allocating computer resources in a virtual computer system, comprising: 7. The virtual computer monitor according to claim 6, wherein during the movement of the main storage area of the second OS, the virtual machine monitor simulates a start instruction of use of the asynchronous processing device issued by the second OS. And a step of analyzing which area of the main storage area of the second OS is accessed by the asynchronous processing device during the simulation, and the area to be accessed is after the access by the asynchronous processing device is completed, A method for dynamically allocating computer resources in a virtual computer system, comprising the step of moving. 8. A plurality of operating systems on a computer including a central processing unit, a main storage unit, and an asynchronous processing unit that performs processing asynchronously with the central processing unit in response to an instruction from the central processing unit. OS) running in a virtual machine system under the control of a virtual machine monitor, when moving a fixed area of a continuous address of the main memory allocated as a main memory area of any of the OSs, the main memory of the OS Defining a correspondence between an area and the fixed area in a virtual space generated by the virtual machine monitor; and changing a mapping destination of the virtual space to a fixed area of a continuous address of a movement destination on the main storage device. And a step of allocating the fixed area of the migration destination as the main storage area of the OS, in the virtual computer system Dynamic allocation method of resources. 9. The method according to claim 8, further comprising, after the defining step, making the size of the fixed area of the mapping destination of the virtual space smaller than the size of the main storage area of the OS. A method for dynamically allocating computer resources in a virtual computer system.