JP2005267498A - Method and system for installing software - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To solve the problem that in a blade server system, when each server accesses to a media reader to install software, this operation requires excessive work time. <P>SOLUTION: Install information of the media reader 30 is written in local disks of all servers 101 to 10n, and each of the server 101 to 10n uses its local disk to install a program. This enables shortening of the time for occupying the media reading means which is a shared device. In the entire system, install time is shortened, compared with a case where each of the servers 101 to 10n uses the reader 30 to sequentially install the program. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a software installation method and system, and more particularly, to a method and system for efficiently installing software from a medium reader to each server in a configuration in which a plurality of servers share a single medium reader.

  In recent years, there is a server form called a blade server system. This is a function that is equivalent to a server system consisting of a plurality of computers by connecting a plurality of CPU blades, which are equipped with functions that constitute a general computer, such as a CPU, memory, and hard disk to a single circuit board called a CPU blade. This is a form to be realized. In the blade server system, in order to reduce the size of the device and reduce the device cost, only one peripheral device such as a medium reading device used exclusively for software installation is mounted, and a plurality of CPU blades share this peripheral device. Take.

  In the case of a system that shares one peripheral device with a plurality of servers, such as a blade server system, the peripheral device can be used by only one server at a time. For this reason, when installing software on all the servers, the operator needs to perform the installation operation for the number of servers, which causes an increase in system operation management costs.

  On the other hand, as an installation method when there are a large number of computers on which software is to be installed, for example, in Patent Document 1, each installation destination computer copies software from an installation source computer based on predetermined work instruction information. It is disclosed that installation is performed and load is distributed by preventing work from being concentrated on one computer at a time.

Patent Document 2 discloses an automatic installation system that installs programs by transmitting necessary installation programs from a control information processing terminal to individual information terminals via a distributor. .
JP-A-10-105406 Japanese Patent Laid-Open No. 2002-7135

  The problem to be solved by the present invention is that, for example, in a blade server system, it takes too much work time to install software by accessing the medium reader for each server. It is an object of the present invention to provide a method and system for installing in a short time on all servers when an installation medium is loaded on a medium reader.

  The present invention provides a system in which a switching unit that connects a plurality of servers and a medium reading unit transmits the contents of a medium to each server in order, and each server temporarily stores this on a local disk and installs it from there. The main feature is that it significantly reduces the overall installation time.

  Specifically, the method according to the present invention is a software in a server system in which a plurality of servers are connected to a single exchangeable recording medium reader, and switching means for switching the connection between the server and the medium reader is provided. An installation method, wherein the switching device sequentially transmits the contents of the installation medium to each server when the software installation medium is mounted on the reader, and the contents of the installation medium received by each server Are stored in the local disk of the local server, and the software is installed by executing the installation program stored in the local disk in each server.

  When the storage of the contents of the installation medium is completed in one server, the switching means determines the server to use the medium reader next time, and notifies all the servers of the installation medium by notifying the determined server It is characterized by saving the contents of.

  The server is characterized by installing software by executing the program when there is a program of an automatic start setting (Auto Run) in the contents of the installation medium upon completion of storage in the server.

  A storage area in which the contents of the installation medium can be stored in advance is secured by firmware of the server on the local disk of the server.

  A software installation system according to the present invention includes a plurality of servers, a single interchangeable recording medium reading device, and switching means for switching connection between the server and the medium reading device, and the switching device includes the reading device. When the software installation medium is mounted on the device, the contents of the installation medium are sent to each server in turn, and each server stores the contents of the installation medium on its own local disk and then saves it on the local disk. The software is installed by executing the installed program.

  In this software installation system, the switching device includes server management information for managing a server that uses the medium reading device, and the server management information does not use the medium reading device in any server, and When the exchangeable recording medium is mounted on the medium reading device in a state in which none of the server computers has issued a medium change request, the contents of the installation medium are sequentially transmitted to each server computer. It is characterized by that.

  Write the installation information of the media reader on the local disk of all servers, and use each local server to perform program installation to save time for occupying the media reader as a shared device. Thus, the installation time of the entire system can be shortened as compared with the case where each server sequentially installs using a reader.

  In addition, there are a plurality of files required for installation of the normal program, and each of the files at different locations on a replaceable recording medium (for example, CD-ROM) mounted on a medium reader at a necessary work place is individually provided. While it takes time to access and proceed with the installation, the time to read all the contents of the medium at once and copy it to the local disk, and the time to install from the local disk are very short. Therefore, even if the time for copying from the media reader to the local disk is included, installation from the local disk can be shortened compared to installing from the media reader, and this time is reduced by the number of servers. Can be achieved.

  The best mode for carrying out the present invention will be described in detail with reference to the drawings. Hereinafter, a case where the present invention is applied to a blade server system will be described. However, the present invention is not limited to this embodiment, and can be applied to all systems in which a plurality of servers share one medium playback device.

  FIG. 1 is a schematic diagram showing the configuration of a blade server system. The blade server system includes n CPU blades 101 to 10n, one medium reading device 20, and a switching device 30 that controls connection between the CPU blade 10 and the medium reading device 20. Each CPU blade 10 and the switching device 20 are connected via the bus 4, and the medium reading device 30 is connected to the switching device 20.

  FIG. 2 is a schematic diagram showing the configuration of the CPU blade 10. The CPU blade 10 has a CPU 11, a memory 12 (main storage device), and a local disk 13 (auxiliary storage device) mounted on a single circuit board, and has a function as a single computer. On the CPU 11, a firmware 14 that controls system activation and an operating system that operates the system operate.

  As shown in FIG. 1, the switching device 20 is connected to CPU blades 101 to 10 n and includes blade management means 21 and device control means 22 for controlling the medium reading device 30. The blade management means 21 and the apparatus control means 22 have an interface with each other. The medium reader 30 is a reader of a replaceable recording medium such as a CD-ROM, and is under the switching device 20. The medium reader 30 does not have to be dedicated for reproduction, and may have an information writing function and other functions.

  The CPU 11 activates the firmware 14 when controlling the medium reading device 30 and the local disk 13. The firmware 14 includes a device driver unit 15 that controls the medium reading device 30 and a disk driver unit 16 that controls the local disk 13. The device driver unit 15 has an interface with the medium reading device 30.

  The memory 12 stores firmware 14 and operating system program codes and various data, and records system information such as a CPU instruction sequence and operation data. Although data used mainly by the OS is recorded / stored on the local disk 13, a medium image storage area 17 is further secured in the system of the present invention. This area is fixedly allocated by the disk driver unit 16 of the firmware 14 before installation of the operating system. As will be described in detail later, the contents of the recording medium set in the medium reading device 30 are stored in this area 17.

  The operation of this blade server system will be described below. FIG. 3 is a flowchart for explaining the operation of the disk driver unit 16 of the firmware 14 in each CPU blade. When the disk driver unit 16 receives the message, it determines the type of message (step A1). If the message is an initialization request for the local disk 13, the disk driver unit 16 initializes the local disk 13 (step A 9), and then sets the medium image storage area 17 as a firmware dedicated area that cannot be referred to by the operating system. Secure (step A10). This area 17 is secured in a size that can store all the information of the media mounted on the medium reader 30. That is, for example, when the medium reading device 30 is a CD-ROM device, an area for a CD-ROM capacity (for example, 650 MB) determined by the standard is secured.

  If the message is a request for writing a medium image from the device driver unit 15, data sent from the medium reading device 30 via the device driver unit 15 is written into the medium image storage area 17 (step A2). After that, if there is an Auto Run designation file for software installation in the written information, the file is activated (step A3).

  On the other hand, if the message is an execution request for a program in the interchangeable recording medium sent from the device driver unit 15, it is checked whether the specified program exists in the medium image storage area 17 (step A4). In the case (step A5), the program is started (step A6). If the designated program does not exist, an error is returned to the device driver unit 15 (step A7).

  Although not described because it is not a feature of the present invention, the disk driver unit 16 performs general write / read control on the local disk 13.

  Next, the operation of the device driver unit 15 of the firmware 14 will be described. FIG. 4A is a flowchart showing the operation of the device driver unit 15, and FIG. 4B is a content image of the media information 18 used by the device driver unit 15. This media information 18 is registered in the memory 12. The device driver unit 15 has an interface with the blade management unit 21 of the switching device 20 and can control the medium reading device 30 by making a request to the blade management unit 21.

  Upon receiving the message, the device driver unit 15 determines the type of message (step B1). When the message is a medium mount notification from the blade management unit 21, in order to read information from the medium reading device 30, the blade management unit 21 is requested to read (step B5).

  If the message is media data sent from the blade management means 21, the device driver unit 15 registers the media name of the received media data in the media information 18 (step B 2), and sends the sent media data to the local disk 13. The disk driver unit 16 is requested to save the file (step B3). When the storage of the medium data is completed, the device driver unit 15 notifies the blade management unit 21 of the completion of capture (step B4).

  If the message is an ejection notification from the blade management means 21, the device driver unit 15 clears the media name currently registered in the media information 18 (step B6).

  When the message is a program execution request on a medium in the medium reading device 30 sent from the upper software such as an operating system, the device driver unit 15 does not start the designated program but the disk driver 21 Request to the unit 16 (step B8). As a result, the designated program is started from the local disk 13 instead of the medium reader 30.

  Further, if the message is a medium change request sent from the upper software, the device driver unit 15 requests the blade management means 21 to change the medium (step B7).

  Next, the operation of the blade management means 21 of the switching device 20 will be described. FIG. 5A is a flowchart showing the operation of the blade management unit 21, and FIG. 5B is a content image of the blade management information 19 used by the blade management unit 21. The blade management information 19 is registered in the blade management means 21 itself or in the internal memory of the switching device 20. As shown in FIG. 5B, the blade management information 19 includes the CPU blade number 19a for registering the management number of the CPU blade that is using the medium reading device 30, and whether the medium reading device 30 is in use. Status information 19b indicating media name 19c in which the name of the media currently mounted on the media reader 30 is registered, and a media change request 19d indicating whether there is a current media change request.

  In the CPU blade number 19a, “1” to “n” indicating the CPU blade numbers 101 to 10n being used or “0” indicating that no CPU blade is being used is registered. In the status information area 19b, “0 (unused)” indicating that there is no CPU blade currently using the medium reader 30 and “1 ( "In use") is stored. When “1” is stored in the status information area 19b, it indicates that the CPU blade registered in the CPU blade number 19a is currently using the medium reading device 30.

  In the media information area 19c, the media name of the medium mounted on the medium reading device 30 is registered, and “none” is registered when the medium is not mounted. In the medium change request area 19d, “1 (change request is present)” indicating a state where there is a medium change request or “0 (no change request)” indicating a state where there is no request is registered. This blade management information 19 can be updated only by the blade management means 21.

  As shown in FIG. 5A, when the blade management unit 21 receives a message, it determines the message type (step C1). Here, if the message is an interrupt notification due to the medium reading device 30 detecting the medium mounting (C2: Y), the process proceeds to the process shown in FIG.

  FIG. 6 is a flowchart for explaining the operation of the blade management means 21 when a mount interrupt notification is received. The blade management means 21 that has received the mount notification from the medium reading device 30 confirms the medium change request 19d (step E1), and in the case of “0 (no request)”, the media name read from the mounted medium is blade management. It is determined whether it is the same as the media information 19c registered in the information 19 (step E2). If the media is different, the status information 19b of the blade management information 19 is confirmed (step E3). If the status information 19b is “0 (unused)” (step E3: Y), the blade management unit 21 registers “1” indicating the first CPU blade 101 in the CPU blade number area 19a (step E4). The media name of the mounted medium is registered in the media information area 19c (step E5), and the device driver unit 15 of the registered CPU blade 101 is notified that a new medium has been mounted (step E5). E6). As a result, when a new medium is mounted when there is no change request and the medium reader 30 is unused, the head CPU blade 101 is notified of the mount. The device driver unit 15 of the CPU blade notified of the medium mounting returns a medium read request in step B5 in FIG.

  If the medium change request 19d is “1 (change request)” in step E1, the change request 19d is updated to “0 (no change request)”, and the media name is registered and mounted on the CPU blade in use. (Steps E5 to E6). Thereby, when a medium change request is issued by any of the CPU blades, the CPU blade is notified that a new medium has been mounted.

  If the media names are the same in step E2, that is, if the same media is remounted, the blade management unit 21 notifies the CPU blade currently in use of the mount (to step E6). As a result, when the medium reading is interrupted for any reason in any of the CPU blades, the reading process in the same CPU blade is resumed by the remounting operation by the operator.

  In the case where the message is an interrupt notification due to the medium reader 30 detecting the medium ejection (step C3: Y) by the message determination in step C1 in FIG. 5, the blade management unit 21 proceeds to the process shown in FIG. The blade management means 21 refers to the status information 19b of the blade management information 19, and when the status information 19b is “1 (in use)” (step F1: Y), further refers to the medium change request information area 19d (step F2). ). When the medium change request information 19d is “0 (no change request)”, the device driver unit 15 of the CPU blade currently in use is notified that the medium has been ejected (step F3), and the status information 19b is set to “ It is updated to “0 (unused)” (step F4).

  On the other hand, if the medium change request information 19d is “1 (change request is present)” in the determination in step F2, the media information 19c is changed to “none” without notifying the CPU blade 10 of the ejection (step F5). The status information 19b is changed to “0 (unused)” (step F4). If the status information 19b is “1 (unused)” in the determination in step F1, that is, if there is no CPU blade in use, the process is terminated.

  If the result of the message type check in FIG. 5 is that the media data is sent from the media reader 30 (step C4: Y), the blade management means 21 confirms the CPU blade number area 19a of the blade management information 19 Then, the medium data is sent to the device driver unit 15 of the CPU blade currently in use (step C5).

  As a result of the message type check (step C1), if it is a medium change request from the device driver unit 15 of any of the CPU blades (step C7: Y), the process proceeds to the process shown in FIG. In this process, the blade management means 21 refers to the blade management information, and when the state information 19b is “0 (unused)” (step G1), the state information 19b is set to “1 (in use)” (step G4), the medium change request area 19d is set to “1 (change requested)” (step G3). Even if the status information 19b is “1 (in use)”, if the requesting blade is currently using the medium reading device 30 (step G2: Y), the medium change request area 19d is set to “1 ( There is a change request) ”(step G3). When this message is received, there is a medium change by the operator thereafter, and a medium ejecting and mounting interrupt message is sent from the medium control means 22.

  On the other hand, if it is determined in step G2 in FIG. 8 that a CPU blade different from the requesting blade is using the medium reading device 30, the blade management information 19 is not changed, and an error is output to the requesting CPU blade. Reply (Step G5). That is, when another CPU blade is using the medium reading device 30, a change request from another blade is rejected.

  If the result of the message type check (step C1 in FIG. 5) is a notification of acquisition completion from the device driver unit 15 of the CPU blade 10 (step C8: Y), the process proceeds to the process shown in FIG. In this processing, the blade management means 21 refers to the blade management information 19 and the notification source blade is using the medium reading device 30 (step H1: Y), that is, the blade registered in the CPU blade number area. If the number 19a matches the blade number of the notification source, the status information 19b is set to “0” and the CPU blade number 19a is incremented by one (step H2). If the updated CPU blade number 19a is n or less (step H3: N), the blade management unit 21 transmits a medium mount notification to the registered blade in the CPU blade number area 19a (step H4). As a result, when the capture of the medium data is completed by a certain CPU blade, the next blade is notified of the medium mount.

  On the other hand, if the CPU blade number 19a exceeds n by increment (step H3: Y), “0” is stored in the CPU blade number area 19a, the status information 19b is set to “0 (unused)”, and the media information 19c. Is changed to “none” (step H5). Further, the medium control unit 22 is requested to eject the medium (step H6). If a CPU blade different from the requesting source is using the medium reading device 30 in the determination in step H1, the process is terminated without doing anything (step H1: N).

  If the result of checking the message type in step C1 in FIG. 5 is a read request from the device driver unit 15 of any CPU blade (step C9: Y), the processing shifts to the processing shown in FIG. In this process, the blade management means 21 refers to the blade management information 19 and confirms whether the CPU blade number 19a matches the requesting blade number (step I1). If they match, the status information 19b is changed to “1 (in use)” (step I4), and an information read request is sent to the device control means 22 (step I5).

  If the blade numbers do not match (step I1: N), the status information 19b is confirmed (step I2). If “1 (in use)”, an error is returned to the requesting device driver unit 15 (step 1). I6). On the other hand, if none of the CPU blades is using the device 30 and the status information 19b is “0 (unused)”, the CPU blade number 19a in the blade management information 19 is changed to the requesting CPU blade number ( In step I3), the status information 19b is changed to "1 (in use)" (step I4), and the medium reading device 30 is requested to read data (step I5).

  FIG. 11 is a diagram illustrating state transition of the blade management information 19 in an operation example. When none of the CPU blades 10 uses the medium reading device 30 and no medium is mounted on the reading device 30, the state J1 is set. In this state, the CPU blade number 19a is “0 (no CPU blade in use)”, the state information 19b is “0 (unused)”, the media information 19c is “None”, and the medium change request 19d is “0 (change). No request) ”.

  When the software installation medium is mounted on the medium reading device 30 in this state J1, the medium reading is notified from the medium reading device 30 to the blade management means 21 via the device control means 22, and as shown in FIG. “1 (shows CPU blade 101)” is registered in the CPU blade number area 19a (step E4), and the mounted media name (here, “A”) is recorded in the media information 19c (step E5). Then, a message notifying that the medium is mounted is sent to the CPU blade 101 (step E5). This state is a state J2 shown in FIG.

  The device driver unit 15 of the firmware 14 operating on the CPU blade 101 receives this and returns a read request to the blade management means 21 in order to store the contents of the mounted medium in the local disk 13 in the own CPU blade ( FIG. 4 step B5). The blade management means 21 that has received the read request confirms that the blade number of the request source and the blade number 19a of the blade management information 19 match in the blade management information 19, and then sets the status information to “1 (use Middle) ", and a read request is made to the apparatus control means 22 (processing in FIG. 10). This state is a state J3 shown in FIG.

  The medium data read by the device control means 22 is sent to the requesting CPU blade 101 via the blade management means 21 (step C5 in FIG. 5). The device driver unit 15 of the CPU blade 101 requests the disk driver unit 16 to store all received medium data in the medium image storage area 17 of the local disk 13. In response, a capture completion notification is transmitted (steps B2 to B4 in FIG. 4). In response to this, the blade control means 21 confirms that the blade number of the notification source matches the CPU blade number of the blade management information 19 and then changes the status information to “0 (unused)”. At the same time, the blade number area is incremented by 1, and a mount message is sent to the next CPU blade 102 (process of FIG. 9). The blade management information at this time is in the state J4 shown in FIG. If the read request for the mount notification to the CPU blade does not wait for a predetermined period, the same processing is performed to update the blade management information to the state J4 and send a message to the next blade. When the blade number after the increment is smaller than n, this is repeated, and the medium images are sequentially taken into the CPU blade 101, the CPU blade 102,. If there is an auto run setting file in the media data, it is activated and the software is automatically installed in each CPU blade.

  When the blade number 19a of the blade management information 19 exceeds n, all the CPU blades are notified of the mounting of the medium. Therefore, the CPU blade number 19a of the blade management information 19 is set to “0” and the status information 19b. Is changed to “0 (unused)”, the media information is changed to “None”, and an eject request is issued to the medium reading device 30 (processing in steps H5 to H6 in FIG. 9). This state is a state J5 shown in FIG.

  When a medium is physically ejected from the apparatus 30 during operation, the blade management information 19 becomes the states J6 and J7 in FIG. 11 by the processing shown in FIG.

  FIGS. 12A and 12B show the state transition of the blade management information 19 when there is a read request from the device driver unit 15 to the blade management means 21 (the process of FIG. 10). FIG. 12A shows a case where the state information 19b is “0 (unused)” even when the requesting blade does not match the registered blade (state K1). In this case, the CPU blade number 19a is the requesting blade. The blade number x is changed, and the state information 19b becomes “1 (in use)” (state K2).

  As shown in states K3 to K4 in FIG. 12B, if the CPU blade number being used matches the requesting blade number when a read request is received, the medium reading device 30 is requested to read data. On the other hand, if the blade number being used and the requesting blade number do not match, an error is returned to the requesting CPU blade without updating the blade management information 19. At this time, the operation is the same whether the media information 19c is “None” or the mounted medium name is registered.

  FIG. 12C is a diagram illustrating a state transition of the blade management information 19 when a mount notification is received from the medium reading device 30. When the status information 19b is “0 (unused)” and the medium “A” is registered in the media information 19c and a mount notification is received (state K5), the blade management unit 21 remounts the same medium. If it is, the CPU blade number 19a is not changed and the CPU blade is notified of the mount (state K6). On the other hand, when a different medium “B” is mounted, the CPU blade number 19a is changed to “1” and the CPU blade 101 is notified of the mounting. As a result, the media images are stored in order from the top CPU blade 101.

  FIG. 13 is a diagram showing a state transition of the blade management information 19 when a medium change request is received from the device driver unit 15 of the CPU blade 10. Upon receiving a medium change request from the device driver unit 15, the blade management unit 21 confirms the status information 19 b based on the blade management information 19, and if it is “0 (unused)” or “1 (used)” If the requesting blade is still in use, the medium change request is set to “1 (change request present)” (processing in FIG. 8). This state is state L1 in FIG. In this case, the medium is ejected by operator intervention, and another medium is then mounted again.

  When the medium is ejected by the operator's replacement work, a medium ejection interrupt notification is sent from the medium reproducing device 30 to the blade management means 21 via the medium control means 22. The blade management unit 21 updates the status information 19b of the blade management information 19 to “0 (unused)” and the media information 19c to “None” (the process in FIG. 7). The blade management information 19 at this time is in the state L2. In this case, since the medium change request 19d is “1 (change request exists)”, the blade in use is not notified of the ejection.

  Thereafter, when a new medium is mounted, a medium mount notification is received from the medium reproducing device 30 via the medium control means 22. Since the medium change request 19d is “1 (change request present)”, the blade management unit 21 updates the change request to “0” without changing the CPU blade number, and registers the media name 19c (FIG. 6). Processing). At this time, even if the newly mounted medium is the same as the previous one, it is not checked here, the medium name is registered in the medium name 19c and notified to the CPU blade, and a read request is requested in response to this. Is received, the status information 19b is set to “1 (in use)”. The state where the medium “A” is mounted is L3, and the state where the medium “B” is mounted is L4.

  FIG. 14 is a diagram showing a state transition of the media information 18 used by the device driver unit 15 shown in FIG. The media information 18 is cleared when the CPU blade 10 is started up or when there is a medium change request from the upper software (state M1). If the medium is changed, the media information is read and the media information 19c of the blade management information is changed (states M2 and M3).

  FIG. 15 is a flowchart showing the operation of the device control means 22 in the switching device 20. When receiving the message (step N1), the device control means 22 checks the message content (step N2). If the message is a mount or eject interrupt notification from the medium reader 30, an interrupt is transmitted to the blade management means 21 (step N3). If the message is a read request from the blade management means 21, the medium reading device 30 is controlled to read information from the mounted medium (step N4) and sent to the blade management means 21 (step N5). If the message is an eject request from the blade management means 21, the medium is ejected from the medium reader 30 (step N6).

  Next, an operation when installing a program stored in a medium on the CPU blades 101 to 10n will be described. The disk driver 15 of each of the CPU blades 101 to 10n reserves a medium image storage area 17 on the local disk 13 during the initialization process of the local disk 13 before the installation of a program such as an operating system. In this case, the status information 19b of the blade management information 19 is “0”, the media name 19c is “None”, and the change request is “0” (state J1 in FIG. 11).

  In this state, when the program installation medium is mounted on the medium reader 30, an interrupt notification is sent to the medium controller 22 of the switching device 20, and the medium controller 22 notifies the blade manager 21 (step N3 in FIG. 15). ). Upon receiving the interrupt notification, the blade management means 21 refers to the blade management information 19, registers “1 (indicating CPU blade 101)” as the blade number 19 a, registers the medium name as the media name 19 c, and registers the CPU name in the CPU blade 101. A mount notification is transmitted (processing in FIG. 6). The device driver unit 15 of the CPU blade 101 returns a medium read request to the blade management unit 21 (step B5 in FIG. 4). The blade management means 21 that has received the read request checks whether the blade number 19a registered in the blade management information 19 matches the blade number of the request source, and sets the status information 19b to “1 (in use)”. A request to read the medium is made to the apparatus control means 22 (the process in FIG. 10).

  The device control means 22 that has received the read request from the blade management means 21 reads the medium storage information and sends it to the blade management means 21 (steps N4 to N5 in FIG. 15). The blade management means 21 that has received the medium information refers to the blade management information 19 and transmits the medium information to the registered CPU blade 101 (step C5 in FIG. 5). The device driver unit 15 of the CPU blade 101 receives this, registers the medium name in the media information 18 managed by itself, and requests the disk driver unit 16 to write the medium information to the local disk 13 (steps B2 to B2 in FIG. 4). B3). The disk driver unit 16 stores the received medium information in the medium image storage area 17 of the local disk 13. At this time, if there is a program set as Auto Run in the written medium information, the program is started (steps A4 to A7 in FIG. 3). After the writing is completed, the device driver unit 15 transmits a capture completion notification to the blade management unit 21 (step B5 in FIG. 4).

  Receiving this notification, the blade management means 21 refers to the blade management information 19 to confirm whether the blade being used matches the notification source blade, and increments the blade number 19a by one. If the CPU blade number 19a does not exceed n, a mount notification is sent to the next CPU blade. (Process of FIG. 9). As a result, the notified next CPU blade 102 stores the medium information in the local disk 13 in the same manner as the blade 101, and returns a capture completion notification. In this way, the blade management means 21 determines the next CPU blade to use the medium reader 30 while referring to the blade management information 19, and notifies the blades 101 to 10n in order of the media mount (steps H1 to H1 in FIG. 9). H4 treatment).

  When a take-in completion notification is sent from the last CPU blade 10n, the blade management means 21 increments the CPU blade number 19a of the blade management information 19 by one (n + 1), which exceeds n. In this case, it is determined that the media installation for all the CPU blades has been completed, and the blade management unit 21 changes the blade number 19a to “0”, the status information 19b to “0”, and the media information 19c to “None”. Then, the apparatus control unit 22 is requested to eject the medium (steps H5 to H6 in FIG. 9). Upon receiving this request, the apparatus control means 22 performs medium ejection from the medium reading apparatus 30 (step N6 in FIG. 15).

  By writing the installation information of the medium reading device 30 on the local disk of all the servers and performing the program installation using the local disk of each server as described above, the medium reading means which is a shared device Can be shortened, and the installation time of the entire system can be shortened as compared with the case where each server performs the sequential installation using the reader 30.

  In addition, there are a plurality of files that are normally required for installing the program, and each of the files at different locations on a replaceable recording medium (for example, a CD-ROM) mounted on a medium reading device at a place where work is required is individually provided. While it takes time to access and proceed with the installation, the time to read all the contents of the medium at once and copy it to the local disk, and the time to install from the local disk are very short. Therefore, even if the time for copying from the media reader to the local disk is included, installation from the local disk can be shortened compared to installing from the media reader, and this time is reduced by the number of servers. Can be achieved.

  Although the embodiments of the present invention have been described, the technical scope of the present invention is not limited to the above-described embodiments, and can be configured as various other embodiments. In particular, although the above embodiment has been described with a blade server system, the present invention can be applied to all computer systems configured to share a single medium reader with a plurality of server computers connected via a communication network. Further, the configuration of each server and the configuration of the medium reading device are not limited to those described above, and may include various other functions necessary in the implementation environment.

  A program installation method and system according to the present invention is a computer system configured such that a plurality of server computers share one medium reader, and a program is newly installed on each server from a medium mounted on the medium reader. It can be used for the purpose of shortening time.

It is the schematic which shows the structure of the server system of this invention. It is the schematic which shows the structure of CPU blade. It is a flowchart explaining the process of the disk driver part 16 of a CPU blade. It is a flowchart explaining operation | movement of the apparatus driver part 15 of a CPU blade. 4 is a flowchart for explaining the operation of the switching device 20. 10 is a flowchart for explaining the operation when the switching device 20 receives a medium mount notification. 10 is a flowchart for explaining the operation when the switching device 20 receives a medium ejection notification. 10 is a flowchart for explaining the operation when the switching device 20 receives a medium change request. It is a flowchart explaining operation | movement when the switching apparatus 20 receives notification of completion of acquisition. 10 is a flowchart illustrating an operation when the switching device 20 receives a medium read request. It is a figure which shows the state transition of the blade management information 19a. It is a figure which shows the state transition of the blade management information 19a. It is a figure which shows the state transition of the blade management information 19a. It is a figure which shows the state transition of the media information. 4 is a flowchart for explaining the operation of the device control means 22 in the switching device 20.

Explanation of symbols

101 to 10n CPU blade 20 switching device 30 medium reader 11 CPU
12 Memory 13 Local Disk 14 Firmware 15 Device Driver 16 Disk Driver 17 Media Image Storage Area

Claims (6)

  A software installation method in a server system in which a plurality of servers are connected to a single interchangeable recording medium reader, and includes switching means for switching the connection between the server and the medium reader, the switching device comprising: A step of transmitting the contents of the installation medium in order to each server when a software installation medium is mounted on the reader; and a step of storing the contents of the installation medium received by each server in the local disk of the server; A software installation method comprising: executing a program installed in the local disk in each server to install software.   2. The software installation method according to claim 1, wherein when the storage of the contents of the installation medium is completed in one server, the switching unit determines a server to use the medium reading device next and notifies the determined server. A software installation method characterized in that the contents of the installation medium are stored in all servers.   3. The software installation method according to claim 1, wherein when the server has a program with an auto start setting in the contents of the installation medium upon completion of storage in the server, the program is executed by executing the program. A software installation method characterized by installing the software.   4. The software installation method according to claim 1, wherein a storage area capable of storing all of the contents of the installation medium is secured in advance on the local disk of the server by the firmware of the server. Software installation method.   A plurality of servers, a reading device for a replaceable recording medium, and switching means for switching the connection between the server and the medium reading device, wherein the switching device has a software installation medium mounted on the reading device. In this case, the contents of the installation medium are transmitted to each server in turn, and each server stores the contents of the installation medium on the local disk of its own server and then executes the installation program stored on the local disk to execute the software. Software installation system characterized by installing   6. The software installation system according to claim 5, wherein the switching unit includes server management information for managing a server that uses the medium reader, and the server management information uses the medium reader for any server. In the state where no server has issued a medium change request, and when a replaceable recording medium is mounted on the medium reader, each server in turn installs the installation medium. A software installation system characterized by transmitting contents.

Images