Patent Application
20120016971
This application relates to and claims the benefit of priority from Japanese Patent Application No. 2010-158822 filed on Jul. 13, 2010, the entire disclosure of which is incorporated herein by reference.
A server system that is provided with a plurality of servers (hosts), a plurality of I/O (Input/Output) devices, and an I/O switch that is coupled to the servers (hosts) and the devices is known (see Patent Literature 1 for instance). [Patent Literature 1] Japanese Patent Application Laid-Open Publication No. 2007-280237
An I/O switch that is configured for a communication with a standard equivalent to that of a communication inside a server is known as an I/O switch. More specifically, in the case in which a standard of a communication inside a server is PCIe (PCI-Express), a PCIe switch can be adopted as an I/O switch.
A server system that is provided with a plurality of PCIe switches can be constructed. More specifically, a PCIe switch to which a plurality of I/O devices can be coupled (hereafter referred to as a post-stage PCIe switch) can be coupled to a PCIe switch to which a plurality of servers can be coupled (hereafter referred to as a pre-stage PCIe switch) for instance. A plurality of post-stage PCIe switches can also be coupled to one pre-stage PCIe switch. Moreover, a plurality of pre-stage PCIe switches can also be coupled to one post-stage PCIe switch.
In accordance with the above server system, a manager can couple any pre-stage PCIe switch and any post-stage PCIe switch to each other and can construct any I/O configuration (such as a configuration of a combination of a server and an I/O device and a configuration of a path that couple a server and an I/O device to each other) depending on a configuration to a PCIe switch.
However, a knowledge that is associated with an internal structure of a PCIe switch is required to understand an I/O configuration that is desired by a manager and a configuration of a PCIe switch for implement the I/O configuration. Consequently, a construction of an I/O configuration is a work with a high degree of difficulty.
This problem can exist even in the case in which an I/O switch other than a PCIe switch is used.
An object of the present invention is to easily construct an I/O configuration of a server system.
At least one second I/O switch which a plurality of I/O devices is coupled to at least one first I/O switch to which a plurality of servers is coupled. In that case, a configuration that is provided with a symmetric property is constructed. More specifically, a controller can implement a configuration in which all servers that are coupled to the first I/O switch can be coupled to all I/O cards that are coupled to the at least one second I/O switch that is coupled to the first I/O switch.
By constructing a configuration that is provided with a symmetric property, a management system that is coupled to the controller can receive a specification of a server unit (a unit that is configured by at least one server) and an I/O device unit (a unit that is configured by at least one I/O device) that is to be coupled to the server unit. The controller receives the information that indicates a server unit that is specified by the management system and an I/O device unit that is specified by the management system. The controller executes a setting in which at least one I/O device that configures the specified I/O device unit is allocated to at least one server that configures the specified server unit to the second I/O switch to which the at least one I/O device is coupled.
By the above configuration, an I/O configuration of a server system can be easily constructed.
An embodiment (example) of the present invention will be described below in detail with reference to the drawings.
In the following embodiment, a server is a server blade that is a server in a blade shape. An I/O device is an I/O card that is an I/O interface device in a card shape (for instance, an HBA (Host Bus Adapter) or a NIC (Network Interface Card)). A standard of a communication inside a server is a PCIe, and an I/O switch is a PCIe switch.
In the following descriptions, the processing will be described while a “program” is handled as a subject in some cases. In the case in which the program is executed by a processor (for instance, a CPU (Central Processing Unit)), the processor executes the predetermined processing by using a storage resource (such as a memory) and/or a communication interface apparatus (such as a communication port) as it decides proper. Consequently, a subject of a processing can also be a processor. The processing that is described while a program is handled as a subject can also be a processing that is executed by a processor or a controller that is provided with the processor. Moreover, the controller can be a CPU itself or can include a hardware circuit that executes a part or a whole of a processing that is executed by a processor. A program can be installed from a program source to each of the controllers. The program source can be a program distribution server or a storage medium for instance.
A management computer can also be at least one computer. For instance more specifically, in the case in which a management computer indicates the information or a management computer transmits the information for an indication to a remote computer, a management computer is a management system. Moreover, in the case in which a function equivalent to a management computer is implemented by using a plurality of computers for instance, the plurality of computers is a management system (the plurality of computers can include a computer for an indication in the case in which a computer for an indication executes an indication). In the following embodiment, a management computer is a management system.
In the following descriptions, while a number is adopted as a type of the identification information of a variety of targets (such as a slot, a server blade, an I/O card, a port, and a virtual switch), the identification information of other types (such as an alphabetic character, a numeric character, another symbol, and a combination thereof) can also be adopted.
There are at least one server rack 111 and at least one I/O expansion rack 121.
One server rack 111 is provided with at least one server chassis 113. One server chassis 113 is provided with at least one partition 115. One partition 115 is provided with K slots 103 (K is an integer number equal to or larger than 2, K=4 for instance). A server blade 105 is inserted into the slot 103.
A front plane 117 is disposed in one partition 115. The front plane 117 is a circuit board that is coupled to the front side of each of the server blades 105 that have been mounted on the partition 115. The front plane 117 is a circuit board that is configured to implement an SMP (Symmetric Multi Processor) connection. The SMP connection will be described in detail later.
The I/O expansion rack 121 is provided with at least one I/O expansion chassis 123. The I/O expansion chassis 123 is provided with J slots 133 (J is an integer number equal to or larger than 1, J=16 for instance). An I/O card 135 is inserted into the slot 133.
A plurality of servers can also be disposed. A user and a server system can be allocated in which the user of a plurality of users (hereafter referred to as a tenant) can use the at least one server system of a plurality of server systems. The server system is provided with at least two partitions 115 and at least one I/O card 135. The server system does not always include a mechanism that is configured to implement an SMP connection. In that case, one server blade can be handled as one partition.
One server system will be described as an example in the following.
The server chassis 113 is provided with a PCIe switch 201 and a controller 203. The I/O expansion chassis 123 is also provided with a PCIe switch 211 and a controller 213. The PCIe switch and the controller in the server chassis 113 will be hereafter referred to as a pre-stage PCIe switch and a pre-stage controller, and the PCIe switch and the controller in the I/O expansion chassis 123 will be hereafter referred to as a post-stage PCIe switch and a post-stage controller.
The pre-stage controller 203 in each of the server chassis 113 and the pre-stage controller 213 in each of the I/O expansion chassis 123 are coupled to a LAN (Local Area Network) 251. A management computer 261 is coupled to the LAN 251. The management computer 261 is a computer that is used by a manager. Although it is not shown, the management computer 261 is provided with a communication interface apparatus, a storage resource (such as a memory), an input device (such as a keyboard and/or a pointing device), a display apparatus, and a processor that is coupled to them for instance. An apparatus in which an input device and a display apparatus are configured in an integrated manner (such as a display apparatus of a touch panel type) can also be adopted. A communication network of other type can also be adopted as substitute for or in addition to a LAN. The LAN 251 is configured by at least one LAN switch for instance.
The pre-stage controller 203 is coupled to the pre-stage PCIe switch 201. The server blade 105 that has been inserted into the slot 103 is coupled to the pre-stage PCIe switch 201. The pre-stage controller 203 manages the configuration information that includes the configuration that indicates a slot 103 into which the server blade 105 that has been coupled to the pre-stage PCIe switch 201 has been inserted.
The post-stage controller 213 is coupled to the post-stage PCIe switch 211. The I/O card 135 that has been inserted into the slot 133 is coupled to the post-stage PCIe switch 211. The post-stage controller 213 manages the configuration information that includes the configuration that indicates a slot 133 into which the I/O card 135 that has been coupled to the post-stage PCIe switch 211 has been inserted.
All of the post-stage PCIe switches 211 in the server system are coupled to each pre-stage PCIe switch 201 in the server system. Moreover, all of the pre-stage PCIe switches 201 in the server system are coupled to each post-stage PCIe switch 211 in the server system. A plurality of pre-stage PCIe switches 201 can be mounted on one server chassis. A plurality of separate post-stage PCIe switches 211 can be coupled to a plurality of pre-stage PCIe switches 201.
The pre-stage controller 203 is provided with a switch interface (I/F) 402, a MC 403, a storage resource (such as a memory) 404, and a processor 401 that is coupled to them. The pre-stage PCIe switch 201 is coupled to the I/F 402, and the LAN 251 is coupled to the NIC 403.
The storage resource 404 stores the information and a computer program. As the information, the pre-stage configuration information 411 is stored for instance. As a computer program, a blade control program 412 is stored for instance. The processor executes the blade control program 412. The pre-stage configuration information 411 is information that is related to a configuration of the chassis 113 that is provided with the information 411. The information 411 and the program 412 will be described in detail later.
The post-stage controller 213 is provided with a switch interface (I/F) 422, a NIC 423, a storage resource (such as a memory) 424, and a processor 421 that is coupled to them. The post-stage PCIe switch 211 is coupled to the I/F 422, and the LAN 251 is coupled to the NIC 423.
The storage resource 424 stores the information and a computer program. As the information, the setting control information 431 and the post-stage configuration information 435 are stored for instance. As a computer program, a configuration setting program 432, a switch control program 433, and a GUI program 434 are stored for instance. The processor executes the configuration setting program 432 and the switch control program 433. The setting control information 431 is information that is related to a setting of the post-stage PCIe. The post-stage configuration information 435 is information that is related to a configuration of the chassis 123 that is provided with the information 435. The information 431, the information 435, and the programs 432 to 434 will be described in detail later.
At least one post-stage controller 213 is provided with the GUI program 434 in one server system. The GUI program 434 is downloaded to the management computer 261 of the server system and is stored into a storage resource in the management computer 261. The GUI program 434 that has been stored into the storage resource is executed by a processor in the management computer 261. The GUI program 434 can also be downloaded to the management computer 261 from another program source (such as the pre-stage controller 203) as substitute for the post-stage controller 213.
At least one post-stage controller 213 is provided with the configuration setting program 432 in one server system. The configuration setting program 432 communicates with the management computer 261 of a download destination of the GUI program 434. Moreover, the configuration setting program 432 can collect the information that is indicated by the configuration information 435 from all of the post-stage controllers 213 in the server system.
In the present embodiment, the following (X1) to (X3) will be executed:
Each of the processing will be described in the following.
<(X1) Construction of a Basic Configuration that is Provided with a Symmetric Property>
The pre-stage PCIe switch and the post-stage PCIe switch is provided with a plurality of upstream ports, a plurality of downstream ports, and a plurality of virtual switches (hereafter referred to as a VS).
The upstream port is a communication port that is coupled to a device on the upstream side.
The downstream port is a communication port that is coupled to a device on the downstream side.
The VS is a virtual I/O switch, and is provided with a virtual upstream port and at least one virtual downstream port. The VS can add a VHN (Virtual Hierarchy Number) to the input information and output the information with the VHN, or can remove the VHN that has been added to the input information and output the information without the VHN.
A server chassis of a chassis number “n” is referred to as a server chassis #n and a pre-stage PCIe switch in the server chassis #n is referred to as a pre-stage PCIe switch #n in some cases in the following. Similarly, an I/O expansion chassis of a chassis number “m” is referred to as an I/O expansion chassis #m and a post-stage PCIe switch in the I/O expansion chassis #m is referred to as a post-stage PCIe switch #m in some cases in the following. Moreover, a port of a port “p” is referred to as a port #p in some cases in the following. Moreover, a VS of a virtual switch number “q” is referred to as a VS #q in some cases in the following. Moreover, a VS in a pre-stage PCIe switch is referred to as a pre-stage VS, and a VS in a post-stage PCIe switch is referred to as a post-stage VS in some cases in the following. In
An example that is shown in
In such a status, a construction of a basic configuration that is provided with a symmetric property is executed.
The “basic configuration” means a configuration of a path from a server blade to a post-stage VS.
The “basic configuration that is provided with a symmetric property” is a basic configuration in which all of the server blades in the same server system can access all of the I/O cards that are coupled to the specific post-stage PCIe switch. The “specific post-stage PCIe switch” is a post-stage PCIe switch that is coupled to the pre-stage PCIe switch to which the all of the servers are coupled.
The construction of the basic configuration that is provided with a symmetric property is implemented as described in the following for instance.
The pre-stage configuration information 411 (see
As shown in
As shown in
In the above (S3), the configuration setting program 432 executes the processing shown in
The absolute blade # can be calculated based on a first calculating formula in which a chassis #, a partition #, and a blade # in the information that has been collected are variables, or can be calculated by using the correspondence management information 600 shown in
The VS # of the post-stage VS can be calculated based on a second calculating formula in which an absolute blade # is a variable, or can be decided based on the information that indicates a correspondence relationship between a chassis # and an absolute blade # and the VS # of the post-stage VS. For instance, one second calculating formula can be selected from a second calculating formula group that is composed of a plurality of second calculating formulas by using the chassis #. The second calculating formula group is configured by the following formulas for instance:
VS#=4b+5
VS#=4b+6
VS#=4b+7
VS#=4b+8
A value that is added to 4b is any one of 5, 6, 7, and 8. The value is different depending on the second calculating formula group. The second calculating formula group can exist for every post-stage PCIe switch (for every I/O expansion chassis).
Here, “b” is a number of a server blade (a number of a slot into which a server blade has been inserted) and is an integer number in the range of 0 to 7 for instance.
In accordance with an example of
The configuration setting program 432 executes the following decisions creates the blade/VS correspondence information 700 shown in
The switch control program 433 sets the upstream setting information 710 (see
The information 711 and the information 712 of the upstream setting information 710 can be set to the post-stage PCIe switch 211 at the time point when a post-stage upstream port and a post-stage VS that is coupled to the post-stage upstream port are decided (this can be similar for all of the post-stage PCIe switches 211). In this case, the VHN 713 can be added to the upstream setting information 710 based on the above described blade/VS correspondence information 700.
In accordance with the above calculation results (shown in
For the construction of a basic configuration that is provided with a symmetric property, the blade/VS correspondence information 700 (see
A GUI program 434 is installed to the management computer 261. The GUI program 434 is a computer program that has been downloaded from the post-stage controller 213. The GUI program 434 can also be downloaded from a program source of other types (such as a Web server) as substitute for the post-stage controller 213.
The GUI program 434 displays the summary GUI 900. The summary GUI 900 is a GUI that displays the information that is related to the summary of the entire server system that is corresponded to a tenant that is desired by a manager (for instance, a tenant to which a manager belongs). By the GUI 900, a manager can execute a brief survey of the server system. The GUI 900 is displayed in the case in which a login is executed by a manager for instance.
The summary GUI 900 is provided with an entire display region 901 and a physical display region 902 for instance.
The entire display region 901 displays the information that indicates a summary of an operation status of the server system (hereafter referred to as the operation summary information). The operation summary information includes the following information for instance;
The operation summary information is the information based on the operation information that is acquired by the pre-stage controller 203 and the post-stage controller 213 for instance. The operation information that is acquired by the pre-stage controller 203 (the information that is stored into the storage resource 404) includes the information that indicates a power consumption and a temperature for every server blade for instance. The operation information that is acquired by the post-stage controller 213 (the information that is stored into the storage resource 424) includes the information that indicates a power consumption and a temperature for every I/O card for instance.
The physical display region 902 displays the information that indicates an external appearance of the server system (hereafter referred to as the appearance information). The appearance information is based on the following information for instance:
The above (c1) can also be the information based on the information that has been collected from the pre-stage controller 203, and the above (c3) can also be the information based on the information that has been collected from the post-stage controller 213. In
The information that has been collected from the pre-stage controller 203 can include the detailed information that is related to the server blade 105 (for instance, the information that indicates a type of the server blade) for every server blade 105. The information that has been collected from the post-stage controller 213 can include the detailed information that is related to the I/O card 135 (for instance, the information that indicates a type of the I/O card 135) for every I/O card 135. The GUI program 434 can change an appearance that is displayed on the physical display region 902 based on the detailed information. By the above configuration, a manager can identify a chassis and a type of a server blade or an I/O card that is inserted into the chassis by the appearance that is displayed on the physical display region 902. In the case in which an element that is desired by a manager (for instance, a chassis, a server blade, or an I/O card) is specified on the physical display region 902, the GUI program 434 displays the detailed information that is related to the element that has been specified.
The entire display region 901 can display a tenant list. In the tenant list, the information that indicates a plurality of tenants can be displayed in a selectable manner. In the case in which a tenant that is desired by a manager is selected from the tenant list, the physical display region 902 can display an appearance of the server system that is corresponded to the tenant that is desired by a manager and can display the information that indicates a summary of an operation status of the server system as the operation summary information. As described above, the summary GUI 900 can be used for a brief survey of the entire of server systems that are corresponded to a plurality of tenants.
The GUI program 434 displays the setting GUI 1100. The setting GUI 1100 is a GUI that is used for allocating an I/O card that is desired by a manager to a partition 115 that is desired by a manager. The GUI 1100 is displayed in the case in which a prescribed operation is executed by the summary GUI 900 for instance.
The setting GUI 1100 is provided with a setting display region 1101, an unallocated display region 1102, and a physical display region 1103 for instance.
The setting display region 1101 displays the following objects:
The unallocated display region 1102 displays at least zero object 1111 and/or 1113 that has not been moved to the setting display region 1101 (that is, in the case in which there is not an unallocated partition and/or an I/O card, the unallocated display region 1102 does not display the objects 1111 and/or 1113).
At least two (or one) partition objects 1111 and at least two (or one) I/O card objects 1113 that are displayed in the unallocated display region 1102 in the first place are the objects 1111 and 1113 that represent all of the partitions 115 and all of the I/O cards 135 in the server system. This is because “the basic configuration that is provided with a symmetric property” has already been constructed, whereby any one of the I/O cards 135 can be allocated to the server blade 105 in any one of the partitions 115.
A display of the object 1111 and/or the object 1113 in the unallocated display region 1102 is controlled based on whether or not the key information is associated with the object 1111 and/or the object 1113 by the GUI program 434. The key information is the information that is associated with the information that indicates a partition or an I/O card that is corresponded to the object 1111 or the object 1113 in the case in which the object 1111 or the object 1113 is moved to the setting display region 1101. For instance, in the case in which a first partition object 1111 that is corresponded to a first partition is moved from the unallocated display region 1102 to the setting display region 1101, the key information “server 1” is associated with the information that indicates the first partition by the GUI program 434. Moreover, in the case in which an I/O card object 113 that is corresponded to a first I/O card is moved from the unallocated display region 1102, the key information “server 1” that is equal to the key information that is associated with the information that indicates the first partition is associated with the information that indicates the first I/O card by the GUI program 434. It can be found that since the key information “server 1” is associated with both of the first partition and the first I/O card, the GUI program 434 must display the first partition object 1111 and the first I/O card object 1113 in the setting display region 1101, and the GUI program 434 must display the first I/O card object 1113 in the first partition object 1111. Moreover, in the case in which the key information is associated with a partition and the key information is not associated with an I/O card, the GUI program 434 must display the object 1111 that is corresponded to a partition in the setting display region 1101. However, it is found that any I/O card object 1113 has not been associated with the object 1111. Moreover, it is found that the GUI program 434 must display the object 1111 and the object 1113 that are corresponded to a partition and an I/O card which the key information has not been associated with in the unallocated display region 1102. The information that indicates an association of the key information with the information that indicates a partition or an I/O card (hereafter referred to as the display control information) can be notified of from the GUI program 434 to the configuration setting program 432, and the display control information can be held by a post-stage controller. The GUI program 434 can receive the display control information from the post-stage controller and control a display to the setting GUI 1100.
The partition object 1111 is provided with a mark 1121 that indicates the number of slots 103 that are included in a partition 115 that is corresponded to the object 1111 and a slot 103 to which a server blade 105 has been inserted. A manager can find that the number of server blades 105 that are included in a partition 115 according to the object 1111 and the mark 1121 that is included in the object 1111. A display of the mark 1121 is controlled based on the server blade management information 500 (see
The physical display region 1103 displays a summary of an appearance configuration of the server system. The “summary of an appearance configuration” can be an abstract summary or a concrete summary as compared with an external appearance that is displayed in the physical display region 902 of the summary GUI 900 for instance.
A display in the physical display region 1103 and an operation of a manager can also be cooperated with each other. For instance, in the case in which a manager specifies arbitrary object 1111 or arbitrary object 1113 (for instance, an object 1111 or an object 1113 that is desired by a manager is touched by using a pointing device such as a mouse or a finger of the manager), the information that indicates a partition 115 that is corresponded to the specified object 1111 (or a chassis 113 or a rack 111 that includes the partition 115) or an I/O card 135 that is corresponded to the specified object 1113 (or a chassis 123 or a rack 121 that includes the I/O card 135) (or the information that indicates an external appearance or a location for instance) can be displayed. More specifically, the partition object 1111 is specified for instance, an image that represents a chassis 113 that includes the partition 115 that is corresponded to the object 1111 (an image part that represents a chassis 113 in an image that represents an appearance configuration of the server system) can be highlighted (for instance, the image can be displayed in a mode different from other chassis (for instance, a different color or a different pattern)). Moreover, in the case in which the I/O card object 1113 is specified for instance, an image of a chassis 123 that includes the I/O card 135 that is corresponded to the object 1113 (for instance, an image that has been drawn in detail in such a manner that a slot into which an I/O card has been inserted can be found) can be displayed.
The detailed GUI 1300 is a GUI in which the detailed information that is related to an element of the server system (for instance, an element that has been specified by a manager). The detailed GUI 1300 is displayed in the case in which an element that is desired by a manager is specified by using the summary GUI 900 or the setting GUI 1100 for instance. The “element” that is described here is a rack 111 or 121, a chassis 113 or 123, a partition 115, a server blade 105, or an I/O card 135 for instance. The “detailed information that is related to an element” is the information that indicates a type of an element and a mounted location for instance.
A wide variety of display controls can be executed for at least one of the summary GUI 900, the setting GUI 1100, and the detailed GUI 1300. A display control is not restricted to the configuration that is shown in the figures. For instance, a plurality of GUI displays can be switched in one window. Every when a new GUI is displayed, a window can be newly created and a GUI can be displayed in the window. Moreover, each of the display regions that are included in the GUI can be a window.
The above descriptions are an explanation of an GUI that is displayed.
A manager can operate the setting GUI 1100 to allocate an I/O card 135 to a partition 115. An example of the operation will be described in the following.
As shown in
As shown in
As shown in
As shown in
As shown in
As described above, a manager moves an object 1111 that is corresponded to the desired partition 115 to the setting display region 1101, and allocates the object 1113 that is corresponded to the desired I/O card 135 to the object 1111. By this operation, the I/O card 135 that is desired by a manger can be allocated to at least one partition 115 that is desired by a manger.
A manager executes a prescribed operation for constructing the server system to which the allocation has been reflected (for instance, selects a character string of “Setting” from a prescribed pull-down menu of the setting GUI 1100). Corresponding to the operation, the information setting is executed.
<(X3) Information Setting that Conforms to a GUI Operation by a Manager>
In order to construct the server system that conforms to an allocation relationship that has been set to the setting GUI 1100, the information that conforms to the allocation relationship is set to the post-stage PCIe switch 211.
More specifically for instance, the GUI program 434 transmits the information that indicates a correspondence relationship that is displayed in the setting GUI 1100 (a correspondence relationship between the partition 115 and the I/O card 135) (hereafter referred to as the allocation relationship information) to the post-stage controller 213 that is provided with the configuration setting program 432. The allocation relationship information includes the following sub information for every correspondence (set) between the partition 115 and the I/O card 135 for instance:
The switch control program 433 sets the information that conforms to the sub information in the allocation relationship information to the post-stage PCIe switch 211. The information that conforms to the sub information is the information that is equal to the downstream setting information 1200 shown in
By executing the above setting, the I/O card 135 is coupled to the virtual downstream port of the post-stage VS as shown in
For which I/O card 135 of at least two (or one) I/O cards 135 that are desired by a manager is allocated to which server blade 105 of at least one server blade 105 in the partition 115 that is desired by a manager, a manager can specify a server blade 105 to be allocated for every I/O card 135 in the partition 115 by using a GUI (not shown), or a server blade 105 in the partition 115 can be automatically decided by the switch control program 433. For instance, a virtual downstream port that is included in a post-stage VS which an I/O card 135 is coupled can be decided based on a number of a server blade 105 in the partition 115 to which the I/O card 135 has been allocated, a number of the I/O card 135, and a number of the virtual downstream port. Moreover, in the case in which there is a plurality of server blades 105 in the partition 115 and a plurality of I/O cards 135 are allocated to the partition 115, all of the I/O cards 135 can be allocated to the server blades 105 of a relative slot number of 0 in the partition 115. Alternatively, a plurality of I/O cards 135 can be distributed to a plurality of server blades 105 in an even manner. For instance, in the case in which there are two server blades 105 and six I/O cards 135, three I/O cards 135 are allocated to each of the server blades 105.
More specifically, in the case in which there are server blades #0 and #1 in the partition 115 in the chassis #0 and four I/O cards #0, #1, #6, and #7 are allocated to the partition 115, the following processing can be executed:
A construction of a server system is completed and the server system can be operated by executing the above (X1) Construction of a basic configuration that is provided with a symmetric property, (X2) Display of a GUI and a GUI operation by a manger, and (X3) Information setting that conforms to a GUI operation by a manager.
In accordance with the present embodiment, by constructing a basic configuration that is provided with a symmetric property, an operation that is necessary for a manager to construct an I/O configuration of the server system is only a specification of a desired partition 115 and a specification of a desired I/O card 135 in essence. More specifically for instance, a partition object and an I/O card object are displayed in the GUI 1100, and a manager can execute a simple operation to associate a desired I/O card object with a desired partition object. In other words, a PCIe switch is hidden to a manager. Consequently, even in the case in which a manager does not know a detailed setting of a PCIe switch, the manager can allocate (couple) an I/O card 135 that is desired by a manager to a server blade 105 in a partition 115 that is desired by a manager.
A flow of an I/O of the information that is executed for the server system has already been known.
In the example shown in
In the example shown in
<Modification of an Allocation of an I/O Card 135 in Accordance with a Configuration Modification of a Partition>
A manager can operate a prescribed GUI (such as a setting GUI 1100) to modify a configuration of a desired partition 115. More specifically, a manager can increase or decrease the number of server blades 105 that configure an SMP group.
As described above, the SMP group is a group of at least one server blade 105 in which an SMP connection has been executed, that is, a group of at least one server blade 105 that configures a scale-up server. By the SMP connection, at least two server blades 105 of a plurality of server blades 105 are combined to be one, and a server group that is configured by the at least two server blades 105 functions as a server blade with a performance higher than that of one server blade. In other words, a scale-up server can be implemented.
The SMP connection is implemented as described in the following for instance.
As shown in
Each of the server blades 105 is provided with a first processor 10A and a second processor 10B (such as a CPU). The first processor 10A is coupled to a first processor 10A of two other server blades 105, and the second processor 10B is coupled to a second processor 10B of two other server blades 105. A path that couples the first processors 10A to each other and/or a path that couples the second processors 10B to each other are connected or disconnected in a logical manner, whereby the number of server blades 105 that configure an SMP group can be increased or decreased.
The SMP connection can utilize the technology that is disclosed in Japanese Patent No. 4484757 or the technology that is disclosed in Japanese Patent Application Laid-Open Publication No. 2010-079467.
<Processing that is Executed in the Case in which the Number of Server Blades 105 that Configure an SMP Group is Decreased>
In the case in which the number of server blades 105 that configure an SMP group is decreased, each of at least one I/O card 135 that has been allocated to at least one server blade 105 that has been decreased is allocated to any one of the server blades 105 that configure an SMP group. For instance, at least one I/O card 135 that has been allocated to at least one server blade 105 that has been decreased can be distributed to at least one server blade 105 that configures an SMP group in an even manner.
More specifically for instance, as shown in
In the case in which the number of server blades 105 that configure an SMP group is increased, the switch control program 433 (or the configuration setting program 432) judges whether or not “x” server blades 105 that are not being operated exist in the first partition 105 that includes the SMP group. For instance, “x” is a value that is obtained by subtracting the number “z” of server blades that configure the SMP group before a configuration modification from the number “y” of server blades that configure the SMP group after a configuration modification (that is, the increasing number of server blades 105). The switch control program 433 (or the configuration setting program 432) can judge whether or not each of the server blades are being operated based on the operation information that is included in the information that is collected from the pre-stage controller 203 (the information that includes the information that indicates whether or not each of server blades in each of the partitions is being operated) for instance.
In the case in which it is judged that “x” server blades 105 that are not being operated exist in the first partition 105, the switch control program 433 (or the configuration setting program 432) adds “x” server blades 105 to the SMP group and allocates at least one I/O card that has been allocated to the SMP group to all of the server blades of the SMP group after a configuration modification in an even manner. For instance, in the case in which the SMP group before a configuration modification is configured by two server blades, four I/O cards are allocated to the two server blades, and the SMP group after a configuration modification is configured by four server blades, the four I/O cards are allocated to the four server blades on a one-to-one basis.
In the case in which it is judged that “x” server blades 105 that are not being operated do not exist in the first partition 105, the switch control program 433 (or the configuration setting program 432) searches other partitions in which the server blades 105 of a number that is equivalent to a number “y” of server blades that configure the SMP group after a configuration modification are not being operated. In the case in which the second partition is found as such a partition, the switch control program 433 (or the configuration setting program 432) allocates all of the I/O cards that have been allocated to the first partition to “y” server blades that are not being operated in the second partition.
More specifically for instance, as shown in
Even in the case in which the number of SMP groups is wanted to be increased to be 2, a server blade of the first partition #0 cannot be added.
A partition that is provided with at least two server blades that are not being operated is searched based on the information that has been collected from all of the pre-stage controllers for instance. As such a partition as shown in
In accordance with the descriptions with reference to
More specifically, in the case in which the program 433 (or the program 432) receives a request of a configuration modification of a partition (a configuration modification of the SMP group) from a manager, the program 433 (or the program 432) judges whether or not the request is an increase or a decrease in the number of server blades that configure the SMP group in the S2301.
In the case in which the result of the judgment of the S2301 is negative (that is, in the case in which the request from a manager indicates a decrease in the number of server blades that configure the SMP group), the program 433 (or the program 432) judges whether or not a configuration modification is possible in the S2302. For instance, in the case in which there are free virtual downstream ports of the number that is equivalent to that of I/O cards that have been allocated to at least one server blade that is to be decreased in at least one post-stage VS that has been allocated to all of the server blades that configure the SMP group after a configuration modification (the server blades after an decrease), the result of the judgment of the S2302 is positive.
In the case in which the result of the judgment of the S2302 is positive, the program 433 (or the program 432) executes a configuration modification in the S2303. In other words, the program 433 (or the program 432) couples an I/O card that has been allocated to at least one server blade that is to be decreased to at least one post-stage VS that has been allocated to all of the server blades that configure the SMP group after a configuration modification (the server blades after an decrease). In this configuration modification, the information 1200 that has been set to the post-stage PCIe switch is updated (see
In the case in which the result of the judgment of the 52302 is negative, the program 433 (or the program 432) displays the information that indicates that a configuration modification is impossible by the GUI program 434 in the S2304.
In the case in which the result of the judgment of the S2301 is positive, the program 433 (or the program 432) judges whether or not a configuration modification is possible in the same partition (the first partition) in the S2311. More specifically, the program 433 (or the program 432) judges whether or not a server blade that is not being operated of the increased number exist in the partition in which a configuration modification is specified (the first partition).
In the case in which the result of the judgment of the S2311 is positive, the program 433 (or the program 432) executes a configuration modification in the S2312. In other words, the program 433 (or the program 432) can modify an allocation destination of an I/O card that is a part of at least one I/O card that has been allocated to all of the server blades that configure the SMP group before a configuration modification from an existing server blade to a server blade that has been newly added to the SMP group. In this configuration modification, the information 1200 that has been set to the post-stage PCIe switch is updated (see
In the case in which the result of the judgment of the 52311 is negative, the program 433 (or the program 432) judges whether or not server blades that is not being operated of the number of blades after a configuration modification exist in other partition in the S2313.
In the case in which the result of the judgment of the S2313 is positive, the program 433 (or the program 432) executes a configuration modification in the S2314. In other words, the program 433 (or the program 432) allocates all of the I/O cards that have been allocated to the SMP group before a configuration modification in the first partition to a server blade that is a member of the SMP group after a configuration modification in other partition that has been found (the second partition). In this configuration modification, the information 1200 that has been set to the post-stage PCIe switch is updated (see
In the case in which the result of the judgment of the S2313 is negative, the program 433 (or the program 432) displays the information that indicates that a configuration modification is impossible by the GUI program 434 in the S2315.
While the preferred embodiments in accordance with the present invention have been described above, the present invention is not restricted to the embodiments, and various changes, modifications, and functional additions can be thus made without departing from the scope of the present invention as a matter of course.
In the above embodiment for instance, a server blade 105 is specified in a unit of a partition 115. However, the server blade 105 can also be specified in other unit (such as a unit of a rack 111 and a unit of a server blade 105) (for instance, an object is in a unit of a server blade). Moreover, an I/O card 135 can also be specified in a prescribed unit (such as a unit of a rack 123 and a unit of a chassis 123) as substitute for that an I/O card 135 is individually specified.
Moreover, an I/O switch of other type can also be adopted as substitute for a PCIe switch.
Moreover, a management system can be provided with a management server that relays a communication between a management computer 261 and a post-stage controller 213 in addition to the management computer 261.
A plurality of server blades 105 of different types can also be coupled to at least one pre-stage PCIe switch 201. Moreover, a plurality of I/O cards 135 of different types can also be coupled to at least one post-stage PCIe switch 211.
Whether or not a server blade or an I/O card is unallocated can be found by whether or not an object is displayed in an unallocated display region 1102. As substitute for that, whether or not a server blade or an I/O card is unallocated can also be indicated by varying a display mode (such as a color and a width of a line) of an object. In this case, it is not necessary that a display region is divided into a setting display region 1101 and an unallocated display region 1102.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2010-158822 | Jul 2010 | JP | national |
