WO2016006072A1

WO2016006072A1 - Management computer and storage system

Info

Publication number: WO2016006072A1
Application number: PCT/JP2014/068345
Authority: WO
Inventors: 渡岡田; 晋広牧; 山本　彰
Original assignee: 株式会社日立製作所
Priority date: 2014-07-09
Filing date: 2014-07-09
Publication date: 2016-01-14

Abstract

The present invention improves the usage efficiency of computer resources in a storage device. A memory stores a change condition for the use of a first computer resource allocated to a first logical partition by a second logical partition. A processor obtains the usage status of multiple computer resources from the storage device and determines, on the basis of the usage status, whether there is a need to increase the amount of computer resources used by the second logical partition. If it is determined that there is a need to increase the amount of computer resources used by the second logical partition, the processor determines whether the first computer resource satisfies the change condition. If it is determined that the first computer resource satisfies the change condition, the processor issues, to the storage device, a change instruction for instructing that the first computer resource is to be used by the second logical partition.

Description

Management computer and storage system

The present invention relates to a management computer and a storage system.

Currently, storage device consolidation, in which storage devices prepared individually for each business, department, or server, is integrated into a large-scale storage device, has become common. By this consolidation, it is possible to efficiently use the storage apparatus such as unification of unused areas that have been conventionally prepared as buffers in each storage apparatus. In addition, since management that has conventionally been required for each storage apparatus can be unified, it is possible to simplify management such that unnecessary work is unnecessary.

Since the consolidation of storage devices is generalized, when a new business is operated, a storage area is generally allocated from a large-scale storage device from the beginning. As a result, in a data center or the like, a multi-tenancy usage pattern in which a plurality of companies, a plurality of departments, etc. share and use a single storage device is not uncommon.

Consolidation of storage devices has advantages such as efficient use of storage devices and easy management, while business that was originally operated using independent storage devices shares a single storage device. Therefore, depending on how the load is applied to the storage device, there is a risk that the data input / output performance will be affected between each business, and the business processing performance will be unduly deteriorated. For such a situation, a technology is known in which one storage device is divided into a plurality of logical partitions and can be managed individually for each logical partition (Patent Document 1). This document discloses a technique for setting a use upper limit of a resource (computer resource) for each logical partition and performing control so as not to use the resource exceeding the upper limit.

US Patent Application Publication No. 2006/0224854

In recent usage forms of multi-tenancy storage devices called clouds, storage devices are used for a wide variety of tasks. For this reason, the load of each logical partition may not be grasped in advance. Also, the load on the storage device due to each job is not uniform in each time zone, and generally increases and decreases.

In the technique disclosed in Patent Document 1, it is necessary to allocate resources according to the load of each business. However, in order to allocate a logical partition to a business in which the load cannot be grasped in advance, excess resources are allocated in anticipation of an increase in the load. Must be assigned to a logical partition. Further, in a business where the load fluctuates depending on the time zone, there is a possibility that many resources that are allocated but not used are generated in a certain time zone. As a result, the resource utilization efficiency of the storage apparatus is lowered.

In order to solve the above problems, a management computer according to an aspect of the present invention includes a memory and a processor connected to the memory and connected to a storage device. The processor issues an allocation instruction for instructing allocation of a plurality of computer resources in the storage apparatus to a plurality of logical partitions, to the storage apparatus. When an I / O instruction is issued from the host computer to the storage apparatus, the storage apparatus selects a logical partition for processing the I / O instruction from the plurality of logical partitions, and the I / O instruction is Processing is performed using computer resources allocated to the selected logical partition. The memory stores a change condition that is a condition for causing the second logical partition to use the first computer resource allocated to the first logical partition. The processor acquires the usage status of the plurality of computer resources from the storage device, and determines whether or not the usage amount of the computer resources by the second logical partition needs to be increased based on the usage status. When it is determined that it is necessary to increase the usage amount of the computer resource by the second logical partition, it is determined whether the first computer resource satisfies the change condition, and the first computer resource is changed. If it is determined that the condition is satisfied, a change instruction that instructs the second logical partition to use the first computer resource is issued to the storage apparatus.

The outline | summary of embodiment of this invention is shown. 1 shows a configuration of a computer system according to an embodiment of the present invention. 2 shows a configuration of a physical storage device 1200. A resource management table is shown. A logical partition management table is shown. A resource allocation management table is shown. A resource change management table is shown. A resource lending status management table is shown. Indicates resource accommodation processing. The storage management process by the management server 1400 is shown. The I / O processing by the physical storage device 1200 is shown. 10 illustrates a logical partition management table according to the second embodiment. 10 shows a resource lending status management table according to the second embodiment. The resource interchange process of Example 2 is shown. The resource allocation recovery time management table is shown. 10 shows a logical partition management table of Embodiment 3. Indicates resource setting processing. The resource interchange process of Example 3 is shown. Indicates failure handling.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. In addition, this embodiment demonstrates the characteristic of this invention, Comprising: This invention is not limited. This embodiment has been described in sufficient detail for those skilled in the art to practice the present invention, but other implementations and configurations are possible without departing from the scope and spirit of the technical idea of the present invention. It is necessary to understand that the configuration and structure can be changed and various elements can be replaced.

Therefore, the following description should not be interpreted as being limited to this. The components of one embodiment can be added to or replaced with the components of another embodiment without departing from the scope of the technical idea of the present invention. As will be described later, the embodiment of the present invention may be implemented by software running on a general-purpose computer, or may be implemented by dedicated hardware or a combination of software and hardware.

In the following description, information used in the present embodiment is mainly described in the “table” format. However, the information does not necessarily have to be represented by a data structure of a table, such as a list, DB, and queue. It may be expressed as a data structure or other.

In the following, when each process in the embodiment of the present invention is described with “program” as the subject (operation subject), the program is executed by the processor and the process determined by the memory and the communication port (communication control device) is performed. It is done while using it. For this reason, the description may be made with the processor as the subject.

Also, the processing disclosed with the program as the subject may be processing performed by a computer such as a management server or a storage device. Part or all of the program may be realized by dedicated hardware or may be modularized.

Information such as programs, tables, and files that realize each function can be stored in non-volatile semiconductor memory, HDD (Hard Disk Drive), storage device such as SSD (Solid State Drive), IC (Integrated Circuit) card, flash memory card , CD (Compact Disc), DVD (Digital Versatile Disc), etc., can be stored in computer-readable non-transitory data storage media, and installed on computers and computer systems using program distribution servers and non-temporary storage media Can do.

Hereinafter, an outline of the present embodiment will be described.

FIG. 1 shows an outline of an embodiment of the present invention.

In the present embodiment, in order to solve the above-described problem, in response to an increase in load that cannot be processed by resources allocated to a certain logical partition, resources allocated to the other logical partition are allocated to the other logical partition. Use when there is no impact on performance. This eliminates the need for excessive resource allocation in preparation for a temporary load increase. As a result, it becomes possible to install more logical partitions in a single storage, and it is possible to improve the utilization efficiency of the storage.

This embodiment is premised on a configuration in which physical storage device resources are divided into logical partitions for use. The resource includes, for example, a disk, a memory, a processor, an FC (Fibre Channel) port, and the like. The resource allocated to the logical partition may be a physical resource unit or a unit obtained by logically dividing the physical resource. The logical partition control function of the physical storage apparatus controls to allocate a specific resource to a specific logical partition and process an I / O instruction for the specific logical partition using the specific resource. The physical storage device is managed by management software on a management server (management computer) connected via a communication network. The management software manages the lending conditions (change conditions) for selecting resources to be lent to other logical partitions from among the resources while guaranteeing the use of the resources allocated to the logical partitions for each resource type. These lending conditions define the usage status of resources before lending, the method of returning resources after lending, and the like.

For example, the management software periodically monitors the load on each logical partition (S1). Here, when it is determined that the memory of the logical partition B is insufficient because the usage rate of the memory allocated to the logical partition B exceeds a certain threshold value (S2), the management software sets the logical partition B to the logical partition B. Based on the memory usage status, it is confirmed whether there is a memory that matches the lending conditions managed by the management software (S3). The lending condition is, for example, that the logical partition of the lending source can immediately lend the memory to the lending destination, and the lending destination can return the memory immediately when the lending source needs the memory. Such lending conditions define, for example, that the lending source uses the memory as a read cache before lending, and that the lending destination uses the memory as a read cache after lending. Here, the fact that the lending source uses the memory as a read cache indicates that the data stored in the memory is also stored in the disk and can be discarded. The use of the memory as a read cache by the lending destination means that when the lending source needs the memory, the data can be immediately discarded and returned to the lending source. When it is determined that the memory that matches the lending condition exists in the logical partition A, the management software issues an instruction to temporarily lend the memory in the logical partition A to the logical partition B to the physical storage device (S4). .

Note that discarding the data stored in the memory whose lending source is the read cache and lending the memory may cause a cache miss for the data and cause a disk read. When the performance requirement of the lender is set in consideration, the performance requirement of the lender can be guaranteed without lending the performance of the lender less than the performance requirement.

As described above, according to the present embodiment, resources can be interchanged between logical partitions while guaranteeing the use of resources allocated to the rented logical partition. The use efficiency of the resources in the storage apparatus is improved by using the resources that are allocated to the logical partition and are not used by other logical partitions.

Hereinafter, examples of the present invention will be described.

FIG. 2 shows the configuration of the computer system according to the embodiment of the present invention.

The computer system includes a host computer 1000, an FC switch 1100, a physical storage device 1200, an IP switch 1300, and a management server 1400.

The host computer 1000 may be a general server or a server having a virtualization function. When the host computer 1000 is a general server, an OS (Operating System) or an application (DB, file system, etc.) running on the host computer 1000 stores data in a storage area provided by the physical storage device 1200. Perform input / output. Further, when the host computer 1000 is a server having a virtualization function, this virtualization function or an application on a VM (Virtual Machine) provided by the virtualization function is stored in a storage area provided by the physical storage device 1200. Input / output data.

The host computer 1000 and the physical storage device 1200 are connected via an FC cable. Using this connection, the host computer 1000 or the VM running on the host computer 1000 inputs / outputs data to / from the storage area provided by the physical storage device 1200. The host computer 1000 and the physical storage device 1200 may be directly connected, but a plurality of host computers 1000 and a plurality of physical storage devices 1200 may be connected via the FC switch 1100. More host computers 1000 and physical storage devices 1200 may be connected by connecting the FC switches 1100 to each other. In this embodiment, the host computer 1000 and the physical storage device 1200 are connected by an FC cable, but when using a protocol such as iSCSI (Internet Small Computer System Interface), they are connected by an IP (Internet Protocol) cable. Alternatively, they may be connected by other connection methods available for data input / output. In that case, instead of the FC switch 1100, an IP switch or a device having a switching function suitable for other connection methods is introduced.

The management server 1400 is connected to the physical storage device 1200 via an IP cable and is a server for managing the physical storage device 1200. The management server 1400 and the physical storage device 1200 may be directly connected, but a plurality of management servers 1400 and a plurality of physical storage devices 1200 may be connected via the IP switch 1300. By connecting the IP switches 1300 to each other, more management servers 1400 and physical storage devices 1200 can be connected. In this embodiment, the management server 1400 and the physical storage device 1200 are connected by an IP cable, but may be connected by other connection methods capable of transmitting and receiving management data. In that case, instead of the IP switch 1300, a device having a switching function suitable for other connection methods is introduced.

The physical storage device 1200 may be further connected to another physical storage device 1200.

The physical storage device 1200 of this embodiment is divided into a plurality of logical partitions 1500 and managed by the management server 1400. The internal configuration of the physical storage device 1200 will be described later.

The management server 1400 includes an input device 1410, an output device 1420, a CPU 1430, a memory 1440, and a NIC (Network Interface Card) 1450. The input device 1410 is a keyboard, a mouse, a tablet, a touch pen, or the like. The output device 1420 is a display, a printer, a speaker, or the like. The CPU 1430 executes processing according to various programs stored in the memory 1440. The memory 1440 is a data storage area such as a RAM (Random Access Memory), and stores various programs, data, temporary data, and the like. Particularly in this embodiment, a logical partition setting program 1441 and logical partition setting management information 1442 are stored. A NIC (Network Interface Card) 1450 is an I / F (Interface) card for connecting an IP cable. When the management network is other than the IP network, the NIC 1450 is an I / F card suitable for the network.

The operation of the CPU 1430 by the logical partition setting program 1441 will be described later.

The number of the host computer 1000, the FC switch 1100, the physical storage device 1200, the IP switch 1300, and the management server 1400 is not limited to the number shown in the figure, and may be any number as long as it is one or more. Further, the management server 1400 may be provided in the physical storage device 1200.

FIG. 3 shows the configuration of the physical storage device 1200.

The physical storage device 1200 includes an FEPK (Front-End Package) 1210, a CMPK (Cache Memory Package) 1220, an MPPK (Microprocessor Package) 1230, a BEPK (Back-End Package) 1240, a disk 1250, and an internal switch 1260.

FEPK1210, CMPK1220, MPPK1230, and BEPK1240 are connected to each other by a high-speed internal bus or the like. This connection may be made via an internal switch 1260.

The FEPK 1210 includes at least one FEIF (Front-End Interface) 1211 that is an interface for data input / output, and is connected to the host computer 1000, another physical storage device 1200, and the FC switch 1100 through the FEIF. When data input / output is performed by communication via an FC cable, an FC port is used. When data input / output is performed by another communication mode, an I / F suitable for the mode is used.

The CMPK 1220 includes one or more cache memories 1221 that are high-speed accessible storage areas such as RAM and SSD. The cache memory 1221 stores temporary data when performing input / output with the host computer 1000, setting information for the physical storage device 1200 to operate various functions, configuration information of the physical storage device 1200, and the like.

The MPPK 1230 includes a NIC 1231, an MP (Microprocessor) 1232, and an LM (Local Memory) 1233.

The NIC 1231 is a management interface and is connected to the management server 1400 and the IP switch 1300. When the management of the physical storage device 1200 is performed by communication via an IP cable, an IP port is used, but when it is performed in a communication mode other than that, an I / F suitable for the mode is used.

The LM 1233 is a high-speed accessible storage area such as a RAM, and stores a program for performing input / output with the host computer 1000, a control program 1234 that is a program for various functions of the physical storage device 1200, and management information 1235 thereof. To do. In particular, the LM 1233 of this embodiment stores logical partition information 1236 for controlling input / output processing and various functions of the physical storage device 1200 according to the set logical partition.

The MP1232 is a processor that executes processing according to a program stored in the LM1233. When a processor that executes processing based on a program stored in the LM 1233 includes a plurality of cores, the MP 1232 may be a core.

The number of NIC1231, MP1232, and LM1233 is not limited to the number described in this figure, and may be any number as long as it is one or more.

The BEPK 1240 includes a BEIF (Back-End Interface) 1241 that is an interface for connecting to the disk 1250. This connection form is generally SCSI (Small Computer System Interface), SATA (Serial Advanced Technology Attachment), SAS (Serial Attached SCSI), etc., but other connection forms may be used.

The disk 1250 is a storage device such as an HDD, SSD, CD drive, or DVD drive.

The number of FEPK1210, CMPK1220, MPPK1230, BEPK1240, disk 1250, and internal switch 1260 is not limited to the number shown in this figure, and may be any number as long as it is one or more.

Here, the function of the control program 1234 will be described.

The control program 1234 includes a data input / output processing program held by a general physical storage device.

The control program 1234 configures a RAID (Redundant Arrays of Inexpensive Disks) group 1270 using a plurality of disks 1250, divides this into logical volumes 1271, which are one or more logical storage areas, and creates a logical volume 1271. It can be provided to the host computer. In this case, the data input / output processing includes processing for converting input / output to / from the logical volume 1271 into input / output to the disk 1250. In this embodiment, it is assumed that data input / output to the logical volume 1271 is performed.

Also, this input / output processing is controlled so that processing is performed using only the resources to which each logical partition 1500 is allocated in order to avoid the influence on the performance of other logical partitions 1500. For example, the processing capacity of the MP1232 is used when the logical partition 1500 performs input / output. When 50% of the usage rate of the MP1232 is allocated to the logical partition 1500, the control program 1234 monitors this usage rate, and when the usage rate reaches the threshold value, the control program 1234 sleeps and processes the other logical partition 1500. Control to give up MP1232 for processing. Alternatively, when 50% of the usage amount of the cache memory 1221 is allocated to the logical partition 1500, the control program 1234 monitors this usage rate, and when the usage rate reaches a threshold, it is used in the logical partition. A part of the cache memory 1221 is released by destage or the like, and after a free area is created, control such as processing is performed.

There are several methods for performing processing using only resources to which the logical partition 1500 is assigned in this way, but it is not specified what method is used in this embodiment. With this control, the processing of each logical partition 1500 is advanced using the allocated resources without being affected by the other logical partitions 1500.

Further, the control program 1234 may have a remote copy function for copying data between two physical storage devices 1200. With this function, the MP 1232 of the copy source physical storage device 1200 reads the data of the copy source logical volume 1271 and transmits the read data to the copy destination physical storage device 1200 via the FEIF 1211. The MP 1232 of the copy destination physical storage device 1200 receives this data via the FEIF 1211 and writes the received data to the copy destination logical volume 1271. In this way, all the data of the copy source logical volume is copied to the copy destination logical volume.

Also, it is necessary to write from the host computer 1000 to the copied area being copied to both the copy source logical volume and the copy destination logical volume. For this reason, the copy source physical storage apparatus 1200 transfers the write command from the host computer 1000 to the copy destination physical storage apparatus 1200. At this time, the write speed may exceed the transfer speed. In this case, the copy source physical storage device 1200 uses the area of the cache memory 1221 as a buffer.

The write from the host computer 1000 to the uncopied area is performed only on the copy source logical volume.

The functions of these physical storage devices 1200 can be variously enhanced and simplified, but the present invention can be applied to these functions without changing the essence thereof. The explanation is based on the premise of the function.

Hereinafter, the logical partition setting management information 1442 stored in the management server 1400 will be described.

The logical partition setting management information 1442 includes a resource management table, a logical partition management table, a resource allocation management table, a resource lending return management table, and a resource lending status management table.

FIG. 4 shows a resource management table.

The resource management table includes an entry for each resource. An entry corresponding to one resource includes a storage apparatus ID 3000, a resource type 3010, a resource ID 3020, and a performance / capacity 3030. The storage device ID 3000 is an ID (identifier) indicating the physical storage device 1200 including the resource. The resource type 3010 indicates the type of the resource. The resource ID 3020 is an ID indicating the entity of the resource. The performance / capacity 3030 indicates the maximum performance or the maximum capacity of the resource.

The value of the resource type 3010 is, for example, “MP_Core” indicating the MP core that is the core of the MP1232, “cache memory” indicating the cache memory 1221, “FEIF” indicating the FEIF 1211, “BEIF” indicating the BEIF 1241, and “ HDD ". The value of the performance / capacity 3030 is, for example, one of the MP1232 core processing speed (MIPS), the capacity of the cache memory 1221 and the disk 1250 (GB), and the performance of the FEIF 1211 and BEIF1241 (Gbps) according to the resource type 3010. is there.

Note that the resource management table may manage resources that are not allocated to logical partitions as unallocated resources.

The logical partition setting program 1441 sets the resource management table based on information input from the administrator and information collected from the physical storage device 1200.

FIG. 5 shows a logical partition management table.

The logical partition management table includes an entry for each logical partition. An entry corresponding to one logical partition 1500 includes a storage device ID 4000, a logical partition ID 4010, a performance requirement 4020, and a load following performance 4030. The storage device ID 4000 is an ID indicating the physical storage device 1200 including the logical partition. The logical partition ID 4010 is an ID indicating the logical partition. The performance requirement 4020 indicates the performance required for the logical partition, and is represented by IOPS (Input / Output Per Second), for example. The load following performance 4030 indicates a time until a resource other than the resource allocated to the logical partition can be used when the load of the logical partition increases. In the load following performance 4030, other resources may be available immediately, or other resources may be available by a set time.

The logical partition setting program 1441 sets the logical partition management table based on a contract with the user of the logical partition 1500 when the administrator creates the logical partition 1500.

FIG. 6 shows a resource allocation management table.

The resource allocation management table includes entries corresponding to combinations of the logical partition 1500 and resources allocated to the logical partition. An entry corresponding to one logical partition 1500 and a combination of resources includes a storage device ID 5000, a logical partition ID 5010, a resource type 5020, a resource ID 5030, an allocation status 5040, and a usage status 5050.

The storage device ID 5000 is an ID indicating the physical storage device 1200 including the logical partition. The logical partition ID 5010 is an ID indicating the logical partition. The resource type 5020 indicates the type of the resource. The value of the resource type 5020 is the same as that of the resource type 3010. The resource ID 5030 is an ID indicating the resource assigned to the logical partition.

Allocation status 5040 indicates the status of allocation to the logical partition among the resources, and indicates a value corresponding to the resource type. When the resource type 5020 indicates MP core, FEIF, or BEIF, the allocation status 5040 indicates an allocation rate that is a ratio that can be used by the logical partition with respect to the maximum performance of the resource. In this embodiment, the cache memory 1221 is managed in units of 4 KB (4096 bytes) blocks. All blocks allocated to the logical partition are registered in the resource allocation management table as resources. When the resource type 5020 indicates a cache memory, the allocation status 5040 indicates the head address of the block that is the resource. When the resource type 5020 indicates a disk, it indicates the capacity that can be used by the logical partition in the resource.

The usage status 5050 indicates the actual usage status of the logical partition among the resources, and indicates a value corresponding to the resource type. When the resource type 5020 indicates an MP core, FEIF, BEIF, or disk, the usage status 5050 indicates a usage rate that is a ratio used by the logical partition with respect to the maximum performance or capacity of the resource. When the resource type 5020 indicates a cache memory, the usage status 5050 indicates the usage (role) of the block. This use indicates what data is stored in the block. For example, the cache memory usage status 5050 is a write cache that stores data before being written to the disk 1250 based on a write command from the host computer 1000 as a cache, or a read that stores data read from the disk 1250 as a cache. Indicates cache etc. The cache memory may be used as a remote copy buffer for storing write data generated during remote copy. At this time, the cache memory usage status 5050 includes a pre-transfer remote copy buffer (pre-transfer RC buffer) storing write data before transfer, and a transferred remote copy buffer storing write data that has been transferred. (Transferred RC Buffer) etc. When the resource is lent to another logical partition 1500, the usage status 5050 indicates a usage rate obtained by adding the rented usage rate to the usage rate used by the logical partition. For example, when the lending logical partition 1500 uses a certain MP core for 10% and the same MP core is lent to another logical partition 1500 for 10%, the usage status 5050 value is 20%. is there. The same applies to the FEIF, BEIF, and disk usage status 5050. When the cache memory is lent out, the cache memory is lent out in units of blocks, so the usage status 5050 indicates the usage of the block at the lent destination. When the resource is unused, the value of the usage status 5050 is “−” indicating that it is unused.

As described above, the resource amount (allocation amount) indicated by the resource allocation state may be an absolute amount such as the MP core processing performance, the FEIF or BEIF bandwidth, the number or size of the cache memory blocks, or the disk capacity. It may be the presence or absence of allocation, or the ratio of the resource amount allocated to the logical partition to the resource amount of the physical resource. Similarly, the resource amount (usage amount) indicated by the use status may be an absolute amount, may or may not be used, may be an application, and is allocated to a logical partition. It may be a ratio of the resource amount used for the logical partition to the resource amount.

The logical partition setting program 1441 sets the resource allocation management table based on information input from the administrator and information collected from the physical storage device 1200 when the administrator creates the logical partition 1500. At this time, the logical partition setting program 1441 issues an allocation instruction for instructing allocation of resources to the logical partition to the physical storage device 1200. Further, the logical partition setting program 1441 updates the usage status 5050 by regular monitoring.

FIG. 7 shows a resource change management table.

Each of the resource lending process and the return process between two logical partitions may be referred to as a change process that changes the logical partition that uses the resource. This table is a table for determining a preparation process that is a pre-process for a change process.

The resource change management table includes an entry for each condition. An entry corresponding to one condition includes a storage apparatus ID 6000, a resource type 6010, a state condition 6020, a preparation process 6030, a processing time 6040, and a necessary resource 6050.

Storage device ID 6000 is an ID indicating the physical storage device 1200. The resource type 6010 indicates the type of resource that can be changed. The state condition 6020 indicates a resource state condition for determining the necessity of the preparation process. For example, the state condition 6020 when the resource type 6010 indicates a cache memory is “when the block before the change process is used as a write cache”. The preparation process 6030 indicates a preparation process performed on the resource when the resource satisfies the condition. When the preparation process is unnecessary, the preparation process 6030 indicates “−”. The processing time 6040 indicates a unit processing time for processing the unit of the preparation process for the time required for the preparation process. For example, when the state of the block before the change process is determined to be a write cache in the state condition 6020, it is necessary to destage the data stored in the write cache as a preparation process. When the preparation process 6030 indicates destaging, the processing time varies depending on the amount of data to be destaged. Therefore, the processing time 6040 is, for example, “5 sec / 100 MB” indicating that it takes 5 seconds for the destage processing of 100 MB data. is there. When the preparation process is unnecessary, the resource can be rented and returned immediately, so the value of the processing time 6040 is “immediate”, for example. The necessary resource 6050 indicates a resource used for the preparation process. When the preparation process 6030 indicates destage and the destage uses the MP core usage rate of 5%, the value of the necessary resource 6050 is, for example, “MP_Core: 5%”.

Here, an example in which a certain block of the cache memory 1221 is used as a write cache has been described. This concept can be applied when the cache memory 1221 is used for another purpose, and can also be applied to other resource types. Another example will be described below.

When a certain block of the cache memory 1221 is used as a read cache, the data stored in the block can be discarded, so that the preparation process is unnecessary. When a certain block of the cache memory 1221 is used as a remote copy buffer and the data stored in the buffer has been transferred, the data can be discarded and no preparation process is required. On the other hand, when the data stored in the buffer is before the transfer, a preparation process for transferring the data to the physical storage device 1200 of the transfer destination is required.

Here, the logical partition that uses the resource before the change process is called the logical partition before the change process, and the logical partition that is used after the change process is called the logical partition after the change process. A case will be described in which the MP core used by the logical partition of the change source is set as the person in charge of I / O processing of a certain logical volume. When the remaining allocation rate, which is the remaining amount obtained by subtracting the allocation rate for the change from the allocation rate of the MP core for the change source logical partition, becomes 0%, the change source logical partition cannot access the logical volume. It is necessary to change the responsibility of the logical volume from the MP core to another MP core. The value of the state condition 6020 indicating this is, for example, “remaining allocation rate = 0%”, and the value of the preparation process 6030 is, for example, “change logical VOL charge”. Furthermore, since the time required for the preparation process depends on the number of logical volumes in charge of the MP core, the processing time 6040 indicates that, for example, it takes 1 second for the change processing in charge of one logical volume. 1 sec / 1 logical VOL ".

When a path to a certain logical volume is set in the FEIF used by the change source logical partition and the remaining allocation rate of the FEIF for the change source logical partition becomes 0%, the change source Since the logical partition cannot access the logical volume, it is necessary to delete the path. The value of the state condition 6020 indicating this is “remaining allocation rate = 0%”, for example, and the value of the preparation process 6030 is “path deletion”, for example. Thereafter, the physical storage device may define a new path to the logical volume using another FEIF.

The logical partition setting program 1441 sets a resource change management table in advance based on information input from the administrator or information collected from the physical storage device 1200.

Resource change management table may be used as load information.

FIG. 8 shows a resource lending status management table.

The resource lending status management table includes an entry for each lent resource. The entry corresponding to one resource includes a storage apparatus ID 7000, a lending source logical partition 7010, a lending resource 7020, a lending destination logical partition 7040, a usage constraint 7050, and a usage status 7060.

Storage device ID 7000 is an ID indicating the physical storage device 1200 including the resource. The lending source logical partition 7010 is an ID indicating the logical partition 1500 that lends the resource. The lending resource 7020 is an ID indicating the resource that is being lent. The lending status 7030 indicates the usage rate that is lent out of the resource and the start address of the block of the resource. The lending destination logical partition 7040 is an ID indicating the logical partition 1500 that borrows the resource. The usage constraint 7050 indicates the usage (specific usage) of the resource by the borrower. For example, the usage constraint 7050 is a borrower constraint that allows the borrower to return the resource immediately or within a specified time when the borrower needs the resource. The usage status 7060 indicates how the borrower is using the resource. For example, if the resource is an MP core, FEIF, BEIF, or disk, the usage status 7060 indicates the usage rate of the resource by the borrower. If the resource is a cache memory, the usage status 7060 stores the type of data stored in the cache memory.

The logical partition setting program 1441 sets the resource lending status management table when lending and borrowing resources and when returning resources. Further, the logical partition setting program 1441 updates the usage status 7060 by regular monitoring.

When the usage status 5050 of a certain resource allocated to a logical partition indicates the usage rate, the value of the usage status 5050 is the amount used in the logical partition, and the loan from the logical partition to another logical partition. It is the sum with the amount that is. For example, when the resource is an MP core, the logical partition to which the MP core is allocated uses 10%, and lends 5% to other logical partitions, the usage status 5050 is 15 %. The same applies to the usage status 5050 of the FEIF 1211, the BEIF 1241, and the disk 1250. When the resource is lent out in the block of the cache memory 1221, the usage status 5050 indicates the usage of the block by the lent destination.

When the logical partition borrows the resource from another logical partition, the usage status 7060 indicates the usage status of the resource by the logical partition. If the logical partition is assigned the resource and borrows the resource from another logical partition, the amount of the resource assigned to the logical partition and the amount of the resource borrowed from the other logical partition Is the total usage status, the value of the usage status 5050 is equal to the value of the allocation status 5040, and the value of the usage status 7060 is a value obtained by subtracting the value of the allocation status 5040 from the total usage status value. . For example, it is assumed that the logical partition is allocated 30% of the MP core, and further 10% of the same MP core is borrowed from another logical partition. When the total usage status value of the MP core by the logical partition is 35%, the usage status 5050 value is 30% and the usage status 7060 value is 5%. When the logical partition is not assigned the resource and the resource is borrowed from another logical partition, the value of the usage status 7060 is the usage rate of the resource by the logical partition.

Hereinafter, the operation of the management server 1400 will be described.

FIG. 9 shows resource accommodation processing.

The logical partition setting program 1441 starts resource interchange processing periodically by the scheduler. When activated, the CPU 1430 acquires the usage status of each resource by each logical partition 1500 from the physical storage device 1200, and based on the acquired usage status, the usage status 5050 of the resource allocation management table and the resource lending status management The table usage status 7060 is updated (S1000).

Next, the CPU 1430 determines whether there is a logical partition that requires the addition of the resource amount (S1010). In this process, the CPU 1430 sequentially selects the selected logical partition and the selected resource type from the physical storage device 1200, and the selected logical partition borrows the resource of the selected resource type from another logical partition based on the resource lending status management table. It is determined whether or not.

When it is determined that the selected logical partition has not borrowed the resource of the selected resource type from another logical partition, the CPU 1430 determines that the usage rate of the usage status 5050 of all the resources of the selected resource type allocated to the selected logical partition is It is determined whether or not the usage rate threshold (upper limit usage amount) has been reached. When the selected resource type is cache memory, the CPU 1430 calculates the ratio of the number of used blocks (the block whose usage status 5050 is not unused) to the number of blocks allocated to the selected logical partition as the usage rate. Further, for example, the CPU 1430 calculates, as the usage rate threshold value, a value obtained by subtracting a predetermined margin from the allocation rate of the allocation status 5040 for each resource. In other words, that the usage rate has reached the usage rate threshold indicates that the usage rate has approached the allocation rate. When it is determined that the usage rate of all resources of the selected resource type assigned to the selected logical partition has reached the usage rate threshold, the CPU 1430 adds the resource amount of the selected resource type to the selected logical partition. Determine that it is necessary.

When it is determined that the selected logical partition has borrowed the resource of the selected resource type from another logical partition, the CPU 1430 determines whether or not the usage rate of the usage status 7060 has reached the usage rate threshold. When it is determined that the usage rate of all resources of the selected resource type borrowed by the selected logical partition has reached the usage rate threshold, the CPU 1430 needs to add the resource amount of the selected resource type to the selected logical partition. (It is necessary to increase the usage amount of the selected resource type by the selected logical partition).

If it is determined as a result of S1010 that there is no logical partition that requires the addition of the resource amount (No), the CPU 1430 ends this flow. On the other hand, if it is determined in step S1010 that there is a logical partition that requires addition of the resource amount (Yes), the CPU 1430 determines the selected logical partition and the selected resource type that are determined to require addition of the resource amount. The additional resource amount, which is specified as the target logical partition and the target resource type, and is the resource amount of the target resource type to be added to the target logical partition is calculated (S1020). The additional resource amount is represented by, for example, a usage rate. A lending unit amount is predetermined for each resource type. The CPU 1430 calculates a value obtained by multiplying the unit amount of the target resource type as the additional resource amount. Note that the additional resource amount may be a fixed value. Further, the CPU 1430 may calculate the additional resource amount by analyzing the resource usage tendency of each logical partition 1500.

After that, the CPU 1430 determines whether there is a rented resource that is a resource of the target resource type that is lent out from the target logical partition to another logical partition based on the resource rent status management table (S1030). If it is determined that there is no rented resource (No), the CPU 1430 advances the process to S1060. On the other hand, if it is determined that there is a lent resource (Yes), the CPU 1430 performs a return process because there is a possibility that a resource shortage can be avoided by returning the lent resource to the lent destination (S1040). In this return process, the CPU 1430 issues a preparation process instruction for instructing a lending destination a preparation process (re-change preparation process) for a lent resource if a preparation process is required, based on the resource change management table. A return instruction (re-change instruction) for instructing the return of the middle resource is issued, and the result is reflected in the resource allocation management table and the resource lending status management table.

Thereafter, the CPU 1430 determines whether or not the additional resource amount has been secured by the return process (S1050). In this processing, the CPU 1430 determines that the additional resource amount has been secured when the resource amount of the target resource type increased by the return processing is equal to or greater than the additional resource amount. If it is determined that the additional resource amount is secured, the CPU 1430 ends this flow. If it is determined that the additional resource amount cannot be secured, the CPU 1430 advances the process to step S1060.

In S1060, the CPU 1430 detects the amount of candidate resources that are candidates for lending among resources of the target resource type allocated to other logical partitions. In this process, first, the CPU 1430 refers to the resource lending status management table, and if another logical partition borrows a resource of the target resource type from another logical partition, is the borrowed resource unused? Determine whether or not. When it is determined that the borrowed resource is in an unused state, the CPU 1430 reflects the result in the resource allocation management table and the resource lending status management table by performing a return process of the resource. As a result, the resource becomes unused. Thereafter, the CPU 1430 refers to the resource allocation management table and calculates the candidate resource amount of each target resource type of the other logical partition. When the target resource type is MP core, FEIF, BEIF, or disk, the candidate resource amount of a certain logical partition is obtained by, for example, subtracting the current usage rate from the resource usage threshold value of the target resource type allocated to the logical partition. Value. When the target resource type is a cache memory, it is the total ratio of the number of blocks serving as candidate resources to the total number of blocks allocated to a certain logical partition. Here, the blocks that are candidate resources include blocks that are not used and blocks that conform to the state condition 6020 of the resource change management table.

Next, the CPU 1430 determines whether or not a lending resource that is a resource satisfying the lending condition is detected from the candidate resources (S1070). In this process, the CPU 1430 selects a logical partition whose candidate resource amount is not 0 as a candidate logical partition, and from the candidate logical partitions, the unused state resource and the preparation process 6030 in the resource change management table preparation process are unnecessary. The indicated resource is detected as a candidate resource, and it is determined whether there is a candidate resource. When it is determined that there is no candidate resource, the CPU 1430 determines that there is no rental resource. If it is determined that there is a candidate resource, the CPU 1430 calculates a return preparation processing time, which is a time required for the preparation processing when the candidate resource is returned, and the load preparation performance 4030 (upper limit) of the logical partition management table is calculated. Determine whether there is a candidate resource that is less than (time). If there is no candidate resource whose return preparation processing time is less than or equal to the load following performance 4030, the CPU 1430 determines that there is no rental resource. If there is a candidate resource whose return preparation processing time is equal to or less than the load following performance 4030, the CPU 1430 determines that the candidate resource whose return preparation processing time is equal to or less than the load following performance 4030 satisfies the lending condition (change condition). Determine the resource as a lending resource.

As a result of S1070, when it is determined that there is no lending resource, the CPU 1430 issues a warning to the administrator indicating that the resource cannot be added to the target logical partition (S1090), and ends this flow.

As a result of S1070, when it is determined that there is a lending resource, the CPU 1430 sets the usage of the resource by the lending destination as a use constraint, and performs lending processing for lending the lending resource from the lending logical partition to the target logical partition (S1080). ), This flow is finished. When a plurality of lending resources are detected in this process, the CPU 1430 performs lending processing by selecting one lending resource from the plurality of lending resources. The CPU 1430 issues a lending instruction (change instruction) instructing to lend a lending resource from the lending source to the target logical partition to the MP 1232 of the lending source and the target logical partition, information indicating the lending resource, and the determined use constraint Are reflected in the resource lending status management table. Based on this lending instruction, the lending source MP 1232 lends a lending resource by performing a preparation process (change preparation process) if necessary, and adjusting the usage rate and usage of the resource. Based on the lending instruction, the lending destination MP 1232 borrows the lending resource by adjusting the usage rate and usage of the resource, and uses the lending resource according to the usage constraint. Further, the CPU 1430 may select the lending resource until the lending resource amount becomes the additional resource amount.

The lending destination uses the lending resource according to the usage constraint, so that the lending destination can return the lending resource at the timing when the lending source needs the lending resource. For example, if the lending resource is a cache memory, the lending destination returns the lending resource immediately when the lending source needs the lending resource by the CPU 1430 setting the lending resource usage constraint in the read cache. Can do.

In this example, the usage status of a resource of a specific resource type is shown as a usage rate. However, if there are resources with different maximum performance in a specific resource type, it is better to manage with the absolute value of the performance. It may be desirable. In this case, the CPU 1430 may manage the usage status of a resource of a specific resource type not by the usage rate but by an absolute value of performance.

In S1070 described above, the CPU 1430 determines that a resource that requires the preparation process cannot be lent, but if the preparation process satisfies the condition, the CPU 1430 may determine the resource as a rented resource. For example, if the block of the cache memory of the logical partition of the lending source is used for the write cache indicated in the state condition 6020, the lending source uses the data indicated in the preparation process 6030 as a preparation process for lending the block. A stage is necessary. For such resources, the CPU 1430 calculates a margin for the usage rate with respect to the allocation rate of the MP core of the rented logical partition, and the margin is indicated in the necessary resource 6050 for the preparation process corresponding to the resource. It is determined whether or not the usage rate is equal to or greater than the usage rate. If the margin is equal to or greater than the required resource 6050, the resource may be determined as a lending resource. Further, the CPU 1430 calculates the required time for the preparation process based on the resource amount and the processing time 6040, and whether or not the required time is equal to or less than the time predetermined by the load following performance 4030 of the target logical partition of the lending destination. If it is determined that the time required for the preparation process is equal to or less than a predetermined time, the CPU 1430 may determine the resource as a lending resource.

In S1010 described above, the CPU 1430 determines that the resource needs to be added to the logical partition when the resource usage rate by the logical partition reaches the usage threshold. As a modified example, the CPU 1430 monitors the performance of the logical partition, and when the resource usage rate by the logical partition reaches the usage rate threshold in S1010, the performance of the logical partition is equal to or higher than the performance requirement 4020 (lower limit performance). If the performance of the logical partition falls below the performance requirement 4020, it is determined that it is necessary to add a resource to the logical partition, and the performance of the logical partition satisfies the performance requirement 4020. If it is, it may be determined that it is not necessary to add a resource to the logical partition. As a result, even if the resource usage rate by a certain logical partition reaches the usage threshold, if the performance of the logical partition is equal to or higher than the performance requirement 4020, it is possible to prevent unnecessary resource lending.

According to the resource accommodation process described above, when the management server 1400 detects a target logical partition that lacks resources, the management server 1400 performs the performance of the lending source logical partition from among the resources allocated to other logical partitions. It is possible to detect a resource satisfying a predetermined condition for the influence on the resource and to allow the resource to be accommodated from the lending logical partition to the target logical partition. When resources in the logical partition of the lending source are insufficient during the lending of resources, the lending source can use the resources within a predetermined time by returning the lending resources. In addition, resources that are allocated to the lending logical partition and are not used are determined as resources that satisfy the lending condition, so that the lending resource can be targeted without affecting the performance of the lending logical partition. It is possible to use the lending resources effectively. In addition, resources that do not require preparation processing associated with lending are determined as resources that satisfy the lending condition, so that lending resources are used in the target logical partition without affecting the performance of the lending logical partition. Can be used effectively. In addition, resources that do not require the preparation process associated with the lending process are determined as resources that satisfy the lending condition, so that the lending resource is used in the target logical partition without affecting the performance of the lending logical partition. be able to. By determining the resource usage status using the state condition 6020, it is possible to select a resource that does not require the preparation process associated with the lending process. For example, if the target resource type is cache memory, the cache memory that stores the discardable data, such as read cache and transferred remote copy buffer, is determined as the cache memory that satisfies the lending condition. The preparation process associated with the process is not necessary. Even if a resource that requires lending preparation processing is used, if the load margin of the logical partition of the lending source is equal to or greater than the resource 6050 required for the resource preparation processing, the resource is determined as a resource that satisfies the lending condition. By doing so, the lending resource can be used by the target logical partition without causing a shortage of resources of the lending logical partition. Further, when a resource whose return preparation processing time is equal to or less than the load follow-up performance 4030 is determined as a resource that satisfies the lending condition, the load-following performance 4030 is obtained when the logical partition of the lending source needs to add a resource. Lending resources can be used within the time indicated in.

FIG. 10 shows storage management processing by the management server 1400.

When the management server 1400 receives a management operation from the administrator, the logical partition setting program 1441 starts storage management processing.

First, the CPU 1430 receives a management operation from the administrator (S2000). Thereafter, the CPU 1430 determines whether or not the management operation conforms to the lending resource usage constraint 7050 (S2010).

When it is determined that the management operation conforms to the usage constraint 7050 (Yes), the CPU 1430 notifies the MP 1232 of a management instruction for executing the management operation (S2020), and this flow is ended. On the other hand, if it is determined that the management operation does not conform to the usage constraint 7050 (No), the CPU 1430 does not execute the management operation and displays a warning indicating that the management operation violates the usage constraint (S2020). End the flow.

According to the above storage management process, the management operation received by the management server 1400 can be determined, so that the usage constraint on the lending resource can be observed. Thereby, the lending destination logical partition can immediately return the lending resource to the lending source logical partition without any preparation process during the return process.

FIG. 11 shows the I / O processing by the physical storage device 1200.

When the MP 1232 of the physical storage device 1200 receives an I / O command from the host computer 1000, the I / O processing is started.

First, the MP1232 receives an I / O command (S3000). Thereafter, when it is necessary to use the lending resource by the processing based on the I / O command, the MP 1232 determines whether the usage of the lending resource by the processing meets the usage constraint 7050 (S3010). When the usage meets the usage constraint (Yes), the MP 1232 processes the I / O command using the lending resource (S3020), and ends this flow. On the other hand, if the usage does not conform to the usage constraint (No), the I / O command is processed without using the lending resource (S3030), and this flow ends.

For example, when the MP 1232 receives a write command from the host computer 1000, it is determined whether or not the rental resource usage constraint 7050 is allowed to be used as a write cache, and the usage constraint 7050 is not allowed to be used as a write cache. For example, a process of using another cache memory as a write cache or a process of writing write data to a disk without using the cache memory is performed.

According to the above I / O processing, the MP 1232 can process the I / O command according to the usage constraint on the lending resource. Thereby, the lending destination logical partition can immediately return the lending resource to the lending source logical partition without any preparation process during the return process.

In the present embodiment, an example is shown in which resource accommodation processing is performed in response to monitoring by the logical partition setting program 1441, but this processing may also be performed by the physical storage device 1200. Further, the trigger for the resource accommodation process may be an instruction from an administrator instead of monitoring, or may be a stage at which an I / O command is received.

In this embodiment, it is assumed that there is no unallocated resource that is a resource that is not allocated to any logical partition in the physical storage device 1200. When there are unallocated resources, the management server 1400 allocates unallocated resources preferentially to the logical partition in which the resource has become insufficient, and at the time when there are no unallocated resources, the management server 1400 allocates resources between the logical partitions. Also good.

According to the present embodiment, it is possible to effectively use the resources of the lending source while guaranteeing the lending source performance by determining whether or not to lend in consideration of the performance of the logical partition of the lending source. Can do. In addition, since the logical partition can be processed by borrowing resources from other logical partitions in response to a temporary increase in load, it is not necessary to allocate excessive resources to the logical partition in advance, and as a result, the physical storage device The operating rate of 1200 resources can be improved.

In the first embodiment, at the time of lending a resource, if the timing at which the lending source logical partition needs the resource (timing to return) is unknown, the resource does not affect the performance of the lending source logical partition. The method of returning is shown. In the present embodiment, a case where a timing at which the lending source logical partition needs the resource is set in advance is shown. In addition, since a present Example becomes the same as Example 1 in many parts, only the difference with Example 1 is shown below.

FIG. 12 shows a logical partition management table of the second embodiment.

Compared with the logical partition management table of the first embodiment, each entry of the logical partition management table of the present embodiment includes a rentable period 8010 and a lentable resource amount 8020 instead of the load following performance 4030. The lending available period 8010 indicates a period during which a resource allocated to the logical partition indicated by the logical partition ID 4010 is permitted to be lent to another logical partition. The rentable resource amount 8020 indicates the amount of resources that can be lent during this period.

The logical partition setting program 1441 sets the logical partition management table based on the input by the administrator when creating the logical partition.

FIG. 13 shows a resource lending status management table of the second embodiment.

Compared with the resource lending status management table of the first embodiment, each entry of the resource lending status management table of the present embodiment further includes a return timing 9010. The return timing 9010 indicates a time at which the lending resource return process is started, and is set so that the return process is completed within the lending available period 8010.

The logical partition setting program 1441 sets a resource lending status management table at the time of lending a resource by resource accommodation processing and when returning the resource.

FIG. 14 shows resource accommodation processing according to the second embodiment.

In the resource accommodation process of the present embodiment, the CPU 1430 performs S1000-S1050 similar to the resource accommodation process of the first embodiment. As a result of S1030, when it is determined that there is no rented resource (No), or when it is determined as a result of S1050 that an additional resource amount cannot be secured by the return process (No), the CPU 1430 advances the process to S4010. . In S4010, based on the logical partition management table, the CPU 1430 identifies a logical partition whose rentable period 8010 includes the current time as a candidate logical partition, rentable resource amount 8020 of the candidate logical partition, and a resource rent status management table. Based on the information of the candidate logical partition, a candidate resource amount that is a resource amount that can be lent by the target resource type in the candidate logical partition is detected. Thereafter, the CPU 1430 determines whether or not a lending resource is detected from the candidate resources (S4020). In this process, as in S1070, the CPU 1430 detects, as candidate resources, resources that are not used and resources that are indicated as not requiring the preparation process in the preparation process 6030 of the resource change management table from the candidate logical partitions. The candidate resource is determined as the lending resource without performing the determination by the load following performance 4030.

As a result of S4020, when it is determined that there is a lending resource (Yes), the CPU 1430 advances the process to S1080. As a result of S4020, when it is determined that there is no rental resource (No), the CPU 1430 advances the process to S1090. Since the processing in S1080 and the processing in S1090 are the same as those in the first embodiment, description thereof is omitted.

After S1080, the CPU 1430 sets a return timer for starting the return process at the return timing 9010 (S4030), and ends this flow. In this process, the CPU 1430 calculates the required time for the return process from the processing time 6040 corresponding to the lent resource in the resource change management table and the amount of the lent resource, and subtracts the required time from the end time of the rentable period 8010. Thus, the return timing 9010 that is the firing time of the return timer is calculated. Since the return process started by firing of the return timer is the same as the content described in S1040, the description is omitted.

According to the above resource accommodation processing, it is possible to reliably return the lending resource within the lending available period 8010 by determining the return processing start timing when lending the lending resource.

According to the present embodiment, the lending resource can be returned before the lending source logical partition actually needs the lending resource. Therefore, the lending source can lend regardless of the lending resource usage at the lending destination. Lending resources can be used later. For this reason, operation becomes possible without the concept like the load follow-up performance 4030 of the first embodiment. This embodiment is particularly effective when a time zone during which an application on a certain logical partition is operated is determined in advance.

In this embodiment, the management server 1400 reserves resources for when a failure occurs in the logical partition. During normal operation, the reserved resources are leased to other logical partitions. In addition, since a present Example becomes the same as Example 1 in many parts, only the difference with Example 1 is shown below.

Compared with the logical partition setting management information 1442 of the first embodiment, the logical partition setting management information 1442 of the present embodiment further includes a resource allocation recovery time management table.

FIG. 15 shows a resource allocation recovery time management table.

The resource allocation recovery time management table includes an entry for each physical storage device 1200. An entry corresponding to one physical storage device 1200 includes a storage device ID 10000 and a failure resource recovery processing time 10010. The storage device ID 10000 is an ID indicating the physical storage device. The failure resource recovery processing time 10010 indicates a target value of time for restoring the resource allocation state to the state before the failure at the time of failure. Since it is difficult to predict in advance the occurrence site of the failure in the physical storage device 1200, it is difficult to guarantee that the lent resource is returned to the lending source at the time of failure. For this reason, it is difficult to guarantee the time for restoring the resource allocation status by exchanging resources in units of logical partitions. Therefore, the management server 1400 looks at the entire physical storage device 1200 and adjusts resources between logical partitions. The failure resource recovery processing time 10010 is a target value of the time for this adjustment.

The logical partition setting management information 1442 according to the present embodiment further includes a reserved resource allocation management table for managing resources reserved for failure in the same manner as the resource allocation management table.

FIG. 16 shows a logical partition management table of the third embodiment.

Compared with the logical partition management table of the first embodiment, each entry of the logical partition management table of the present embodiment includes a resource guarantee level 11010 at the time of failure instead of the load following performance 4030. When a resource allocated to a certain logical partition during normal operation is a normal guaranteed resource amount, and a resource amount guaranteed at the time of a failure is a fault guaranteed resource amount, the resource guarantee level 11010 at the time of failure is Indicates the ratio of the amount of fault guaranteed resources. For example, when the resource guarantee level 11010 at the time of failure of a certain logical partition is 100%, the resource of the normal guaranteed resource amount is reserved for the logical partition, and the resource of the normal guaranteed resource amount is guaranteed at the time of failure. When the resource guarantee level 11010 at the time of failure of a certain logical partition is 50%, the amount of the guaranteed resource is less than the normal guaranteed resource amount, and 50% of the normal guaranteed resource amount is guaranteed at the time of failure. In addition, when the resource guarantee level 11010 at the time of failure of a certain logical partition is lower than 100%, even if there is no resource reserved for the logical partition, by suppressing the resource usage rate during normal operation, The amount of resources that can be used at the time of a failure may be prevented from decreasing compared to the amount of resources that were used during normal operation.

FIG. 17 shows the resource setting process.

When the administrator inputs a management operation for allocating resources to the logical partition to the management server 1400, the logical partition setting program 1441 starts the resource setting process.

First, the CPU 1430 receives from the administrator the target logical partition that is the target logical partition of the management operation and the resource amount to be allocated to the target logical partition (S5000). In this processing, for example, the CPU 1430 may directly accept the normal guaranteed resource amount, which is the resource amount guaranteed during normal operation, and the failure resource guarantee level for the target logical partition. Further, for example, the CPU 1430 may receive the performance requirement at the normal time and the performance requirement at the time of failure for the target logical partition, and calculate the resource amount satisfying each performance requirement.

Next, the CPU 1430 allocates normal resources (third computer resources) corresponding to the normal guaranteed resource amount to the target logical partition, and reflects information indicating the normal resources in the resource allocation management table (S5010). Next, the CPU 1430 similarly allocates a reserve resource (first computer resource) required in the event of a failure to the target logical partition, and reflects information indicating the reserve resource in the reserve resource allocation management table (S5020). . In this process, the CPU 1430 does not have to allocate a reserve resource to the target logical partition when the resource guarantee level at the time of failure of the target logical partition is lower than 100%. Then, the CPU 1430 sets the normal resource and reserve resources in the physical storage device 1200 (S5030), and ends this flow.

According to the above resource setting process, the normal resource and the reserve resource can be allocated to each logical partition in accordance with an instruction from the administrator.

FIG. 18 illustrates resource accommodation processing according to the third embodiment.

In the resource accommodation process of the present embodiment, the CPU 1430 performs S1000 to S1020 similar to the resource accommodation process of the first embodiment. After S1020, the CPU 1430 searches for a reserve candidate resource that is a reserve resource of the target resource type from the reserve resource allocation information management table and the resource lending status management table (S6010). Thereafter, the CPU 1430 determines whether or not a lending resource is detected from the reserve candidate resources (S6020). For example, the CPU 1430 calculates the required time for the reserve candidate resource return process, and if the required time is within the failure resource recovery process time 10010 in the resource allocation recovery time management table, the reserve candidate resource is lent. Determine as.

When it is determined that there is no rental resource (No), the CPU 1430 issues a warning indicating that there is no rental resource (S1090), and ends this flow. On the other hand, when it is determined that there is a lending resource (Yes), the CPU 1430 performs a lending process similar to that in the first embodiment (S1080), and ends this flow.

According to the above resource accommodation processing, if it is determined that a resource needs to be added to a certain logical partition, the resources for reservation of other logical partitions can be lent to the logical partition. During normal operation, the lending source does not use the reserve resource, so the reserve resource can be effectively used without affecting the performance of the lender. By determining the time required for the return process at the time of lending, the lending resource can be returned to the lending source within a predetermined time from the occurrence of the failure.

FIG. 19 shows failure processing.

When the logical partition setting program 1441 detects a failure, the logical partition setting program 1441 executes failure processing. Note that the logical partition setting program 1441 may activate the failure processing according to an explicit instruction from the administrator.

First, when this process is activated, the CPU 1430 performs a lending resource return process for each logical partition that borrows the lending resource (S8000). In this process, the CPU 1430 calculates the total of normal normal resources and borrowed lending resources based on the allocation status of normal resources in the resource allocation management table and the resource guarantee level 11010 at the time of failure, and from the total Calculate the return resource amount by reducing the reserve resource amount. Thereafter, the CPU 1430 issues a return instruction (recovery instruction) for returning the resources corresponding to the return resource amount from the lending destination to the lending source, and manages the returned resources as unallocated resources.

When the return processing of all the logical partitions is completed, the CPU 1430 allocates unallocated resources to the logical partitions so as to satisfy the resource guarantee level at the time of failure for each logical partition (S8010), and ends this flow.

According to the above fault processing, when a fault occurs, it is possible to guarantee resources at the fault resource guarantee level by reallocating resources according to the fault resource guarantee level.

According to the present embodiment, in order to ensure the reliability of the logical partition, when the reserve resource is secured for the logical partition, the reserve resource is utilized by other logical partitions during normal operation. The operating rate of the entire physical storage device 1200 can be improved.

The computer system of this embodiment may be a combination of the above-described plurality of examples.

The present invention is not limited to the above embodiments, and can be modified in various other forms without departing from the spirit of the present invention.

DESCRIPTION OF SYMBOLS 1000 ... Host computer 1200 ... Physical storage apparatus 1210 ... FEPK 1220 ... CMPK 1230 ... MPPK 1240 ... BEPK 1250 ... Disk 1260 ... Internal switch 1400 ... Management server 1430 ... CPU 1440 ... Memory 1500 ... Logical partition

Claims

Memory,
A processor connected to the memory and connected to a storage device;
With
The processor issues an allocation instruction to the storage apparatus to instruct to allocate a plurality of computer resources in the storage apparatus to a plurality of logical partitions;
When an I / O instruction is issued from the host computer to the storage apparatus, the storage apparatus selects a logical partition for processing the I / O instruction from the plurality of logical partitions, and the I / O instruction is Processed using the computer resources assigned to the selected logical partition;
The memory stores a change condition, which is a condition for causing the second logical partition to use the first computer resource allocated to the first logical partition.
The processor acquires the usage status of the plurality of computer resources from the storage device, and determines whether or not the usage amount of the computer resources by the second logical partition needs to be increased based on the usage status. When it is determined that it is necessary to increase the usage amount of the computer resource by the second logical partition, it is determined whether the first computer resource satisfies the change condition, and the first computer resource is changed. When it is determined that the condition is satisfied, a change instruction that instructs the second logical partition to use the first computer resource is issued to the storage apparatus.
Management computer.
The processor determines whether it is necessary to increase the usage amount of the computer resource by the first logical partition based on the usage state, and needs to increase the usage amount of the computer resource by the first logical partition. If it is determined that the first computer resource is used by the second logical partition, it is determined that the first computer resource is used by the second logical partition. In this case, a re-change instruction for causing the first logical partition to use the first computer resource is issued to the storage device.
The management computer according to claim 1.
The processor determines whether the first computer resource is used for the first logical partition based on the usage state, and if the first computer resource is not used for the first logical partition. If determined, the first computer resource is determined to satisfy the change condition;
The management computer according to claim 2.
The processor determines whether or not the load of the first logical partition increases according to the change instruction, and when it is determined that the load of the first logical partition does not increase according to the change instruction, Determining that the first computer resource satisfies the change condition;
The management computer according to claim 3.
The change condition indicates a specific state of the first computer resource by the first logical partition,
When it is determined that it is necessary to increase the usage amount of the computer resource by the second logical partition, the processor, when the state of the first computer resource by the first logical partition is the specific state, Determining that the load of one logical partition is not increased by the change instruction;
The management computer according to claim 4.
The first computer resource is a cache memory;
The specific state is a state in which discardable data is stored in the cache memory.
The management computer according to claim 5.
The memory stores load information indicating an increase in load of the first logical partition due to the change preparation processing for the change preparation processing necessary for causing the second logical partition to use the first computer resource,
The processor calculates a load margin of the first logical partition based on the usage state, determines whether the margin is equal to or greater than the increase amount based on the load information, and the margin is If it is determined that the increase amount is greater than or equal to the increase amount, the first computer resource is determined to satisfy the change condition;
The management computer according to claim 6.
The change instruction indicates a specific use of the first computer resource by the second logical partition,
In response to the change instruction, the first computer resource is used for the specific application from the second logical partition.
The management computer according to claim 1.
In response to the re-change instruction, re-change preparation processing necessary for causing the first logical partition to use the first computer resource used by the second logical partition is performed by the second logical partition. ,
When it is determined that it is necessary to increase the amount of computer resources used by the second logical partition, the processor predicts the time required for the re-change preparation process, and the predicted time is a predetermined upper limit. It is determined whether or not the time is less than or equal to the time, and when the predicted time is determined to be less than or equal to the upper limit time, it is determined that the first computer resource satisfies the change condition.
The management computer according to claim 2.
The allocation instruction indicates that a second computer resource is allocated to the second logical partition;
The processor acquires the performance of the second logical partition measured by the storage device from the storage device,
The usage status indicates a usage amount of the second computer resource by the second logical partition,
The processor sets an upper limit usage amount that is an upper limit of the usage amount based on an allocation amount of the second computer resource to the second logical partition,
When it is determined that the usage amount exceeds the upper limit usage amount, and the performance of the second logical partition is lower than a predetermined lower limit performance, the processor is configured to use a computer resource by the second logical partition. Determine that usage needs to be increased,
The management computer according to claim 1.
The memory stores a period for allowing the second logical partition to use the first computer resource,
If it is determined that the first computer resource satisfies the change condition within the period, the processor sends a change instruction that instructs the second logical partition to use the first computer resource. Issuing to the storage device, and issuing the re-change instruction to the storage device before the end of the period;
The management computer according to claim 1.
The allocation instruction indicates that a third computer resource is allocated to the first logical partition for normal operation, and indicates that the first computer resource is allocated to the first logical partition for failure.
In the state where the first computer resource is used by the second logical partition and the third computer resource is used by the first logical partition in response to the change instruction, the processor When a failure is detected, a recovery instruction for using the first computer resource for the first logical partition is issued to the storage device.
The management computer according to claim 1.
A storage device connected to the host computer;
A management computer connected to the storage device;
With
The management computer issues an allocation instruction to the storage apparatus to instruct to allocate a plurality of computer resources in the storage apparatus to a plurality of logical partitions,
When an I / O instruction is issued from the host computer to the storage apparatus, the storage apparatus selects a logical partition for processing the I / O instruction from the plurality of logical partitions, and selects the logical partition as the selected logical partition. Processing the I / O instruction using the allocated computer resources;
The management computer stores a change condition, which is a condition for causing the second logical partition to use the first computer resource allocated to the first logical partition,
The management computer obtains the usage status of the plurality of computer resources from the storage device, and determines whether it is necessary to increase the usage amount of the computer resource by the second logical partition based on the usage status. When it is determined that it is necessary to increase the usage amount of the computer resource by the second logical partition, it is determined whether or not the first computer resource satisfies the change condition, and the first computer resource is When it is determined that the change condition is satisfied, a change instruction that instructs the second logical partition to use the first computer resource is issued to the storage device.
Storage system.