JP2535640B2 - Arrangement method of server space - Google Patents

Description

The present invention relates to a server space allocation method in which a server space that provides system services to users in a computer system is allocated in consideration of load distribution. For the purpose, in the server space placement method in which a plurality of server spaces, each having a working and a standby, are distributed and arranged on a plurality of processors, the working and standby server spaces to be arranged are sequentially designated. A server space designating unit, a working placement unit that sequentially places the working server spaces sequentially designated by the server space designating unit to adjacent processors, and a count value that holds a count value related to the placement of the standby server space. When the holding means and the active allocating means allocate the active server space, the corresponding standby server space is opened. The standby arrangement for arranging the server empty to the processor shifted by the count value held by the count holding means, and updating the count value every time the processor in the active server space completes one cycle. And means.

[Industrial applications]

The present invention relates to a server space arranging method in which a server space for providing a system service to a user in a computer system is arranged in consideration of load distribution.

[Conventional technology]

A general computer system has various server spaces corresponding to each system service in order to provide various system services to users. Examples of system services include job activation processing and file open / close processing. The server space is a processing unit that performs system services.

In addition, in recent computer systems, it is possible to continue processing even if a part fails.
Some include a plurality of processor modules and have active and standby server spaces. In such a computer system, even if the processor module to which the current server space is allocated becomes inoperable due to a failure, the standby server space allocated to another processor module can provide system services. There is.

Conventionally, the above-mentioned arrangement of each server space in a plurality of processor modules is restricted by the user in advance because there are restrictions such as arranging the active module and the standby module in separate processor modules, and distributing the load to each processor module. It had to be defined, and the computer system was instructed to perform this defined arrangement. The computer system provides various system services by associating each processor module with the active and standby server spaces in accordance with the instructed arrangement.

[Problems to be Solved by the Invention]

By the way, in the above-mentioned conventional method, since the user defines and instructs the arrangement of the server space in consideration of the load distribution and the like, there is a drawback that the burden on the user becomes large. In particular, if the number of processor modules that a computer system has and the number of system services that it provides are different, the way of arrangement that is a combination of them also differs, so it is necessary to define the arrangement for each computer system, and the burden of usage difference is large. Become. For this reason, there has been a demand for a server space layout method that can automate the layout in consideration of load distribution.

The present invention was created in view of the above points, and it is an object of the present invention to provide a server space placement method that enables automation of server space placement.

[Means for solving the problem]

FIG. 1 is a block diagram of the principle of the server space arrangement method of the present invention.

In the figure, a server space designating means 111 in a server space arrangement method in which a plurality of server spaces each having an active and a standby are arranged in a distributed manner among a plurality of processors is used. Specify server space sequentially.

The active allocation unit 121 allocates the active server spaces sequentially designated by the server space designation unit 111 to the adjacent processors in sequence.

The count value holding unit 131 holds the count value regarding the arrangement of the standby server space.

The standby arranging means 141, when the active server space is allocated by the active allocating means 121, holds the corresponding standby server space in the count value holding means 131 for the processor in which the active server space is allocated. It is arranged in the processor shifted by the calculated count value, and the count value is updated every time the processor in which the active server space is arranged completes one cycle.

Therefore, as a whole, the active server space and the standby server space are arranged in the processors that are deviated by the count value.

[Work]

When the active and standby server spaces are sequentially designated by the server space designating means 111, the active placement means 121 places the designated active server spaces in the processors that are connected in sequence, and the standby placement means 141 is designated. The standby server space is sequentially arranged for the processors that are deviated from the current one by the count value. This count value is updated every time the processor that arranges the active server space completes a cycle, so that the standby server space corresponding to the plurality of active server spaces arranged in the same processor is arranged in different processors. It has become.

〔Example〕

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

FIG. 2 shows the configuration of an embodiment to which the server space arrangement method of the present invention is applied.

In the figure, 211 is a processor module (hereinafter referred to as PM) configuration information storage unit, 221 is a server start-up information storage unit, 231 is a PM determination unit, 241 is a server determination unit, 251 is an active placement unit, 261 Indicates a standby arrangement unit, 271 indicates a counter, and 281 indicates an arrangement information storage unit.

The PM configuration information storage unit 211 stores the number of PMs in the target computer system as PM configuration information, and the server startup information storage unit 221 stores a server space prepared for providing various system services. The number and the order of arrangement are stored as server startup information.

The PM determination unit 231 reads the number of PMs stored in the PM configuration information storage unit 211 and performs a process of assigning serial numbers to each PM. The server determination unit 241 reads the number and arrangement order of the server spaces stored in the server activation information storage unit 221, and performs a process of assigning serial numbers to each server space to be arranged. The same serial number is assigned to each of a pair of active and standby server spaces for providing a common system service.

The working placement unit 251 performs a process of placing each working server space, to which a serial number is added by the server determination unit 241, in each PM. On the other hand, the standby arranging unit 261 performs a process of arranging a standby server space corresponding to each active server space allocated by the active arranging unit 251 in each PM. counter
271 is used in this standby arrangement.

The placement information storage unit 281 stores the placement results of the working placement unit 251 and the standby placement unit 261.

Assume that there are four PMs and twelve server spaces in consideration of the takeover process, that is, 12 server spaces each having an active part and a standby part, as the objects of arrangement in the present embodiment.

FIG. 3, FIG. 4 and FIG. 5 show the operation procedure of one embodiment. Further, FIG. 6 shows an outline of the arrangement of the active server space in one embodiment, and FIG. 7 shows an outline of the arrangement of the standby server space in the embodiment. 6 and 7
In the figure, GS (Global Service) indicates each server space, and the numbers assigned to each PM and each server space (hereinafter referred to as GS) are assigned by the PM determination unit 231 or the server determination unit 241. Shows the serial number.

 Hereinafter, FIG. 2 to FIG. 7 will be referred to.

First, the PM determination unit 231 refers to the PM configuration information stored in the PM configuration information storage unit 211 to check how many PMs exist in the computer system in which the PM configuration information is placed (step A1). The number of PMs in this embodiment is 4, and the PM determining unit
231 is a serial number “0” to “3” assigned to each PM.

In addition, the server determination unit 241 has a server startup information storage unit 221.
Refer to the server start information stored in
Check whether to place (step A2). The number of GSs in this embodiment is 12, and the server determining unit 241 assigns serial numbers “1” to “12” common to the active and standby of each GS. The serial numbers indicate the order of GSs that determine the layout, and the layout processing described below is performed sequentially from the GS with the smallest number.

Next, according to the order determined in step A2, the GS that is not arranged and has the highest priority (smallest serial number) is selected (step A3). This selection is performed in each of the active placement unit 251 and the standby placement unit 261 and the active and standby GSs with the same serial number are selected.

Next, the working placement unit 251 places the selected working GS according to a predetermined procedure (step A4). Figure 4 shows
The details of the arrangement process in step A4 will be described below.

First, the active placement unit 251 determines whether the PM in which the active GS is placed one before is the last PM3, that is, the PM3 having the largest serial number (step B1). At the beginning, there is no active GS placed one before, so the active GS is PM
This judgment is made assuming that the card is placed in 3. Therefore, a positive determination is made in step B1.

When the configuration is determined in step B1, the active placement unit 251 places the GS focused on PM0 having the smallest serial number (step B2). On the other hand, if a negative decision is made in step B1, the active placement section 251 places the GS of interest in the next PM (step B3). For example, if PM1 is placed one before, PM2 is placed in step B3. When the placement in step B2 or B3 is completed,
Next, the active placement unit 250 instructs the standby placement unit 261 to place the standby GS (step B4).

By the steps B1 to B4 described above, the working GS is arranged in step A4 of FIG. The placement result of this working GS is stored in the placement information storage unit 281. Fig. 6 shows the arrangement of the active GSs only. As shown in the figure, the active GSs with consecutive serial numbers are sequentially allocated to adjacent PMs, and all PMs are loaded. Are distributed.

Next, the standby placement unit 261 receives the instruction from the active placement unit 251 and places the standby GS selected in step A3 according to a predetermined procedure (step A5). FIG. 5 shows the details of the arrangement processing procedure in step A5, which will be described below.

First, the standby arranging unit 261 determines whether or not the PM on which the active part is arranged is PM0 (step C1). Since GS0 is arranged in PM0 in step B2 described above, an affirmative judgment is first made in step C1.

When an affirmative decision is made in step C1, the standby placement unit 261 adds 1 to the value of the counter 271 and updates the counter 271 (step C2). This counter 271 holds "0" as an initial value, and is updated to "1" by the processing of this step C2.

Next, the standby placement unit 261 determines whether or not the value of the counter 271 is equal to the number of PMs (= 4) (step C3). If a positive judgment is made (if they are equal), the counter 27 is forcibly
The value of 1 is changed to "1" (step C4), and when the determination is negative, the counter 271 is not changed, and then the PM in which the current GS is arranged is shifted by the value of the counter 271. On the other hand, the standby GS that is focused on is placed (step C5).

By the steps C1 to C5 described above, the standby GS is arranged in step A5 in FIG. The placement result of the standby GS is stored in the placement information storage unit 281. FIG. 7 shows a state in which a standby GS is arranged in addition to the arrangement of the working GS shown in FIG. 6, and as shown in FIG.
The standby GSs corresponding to the active GSs assigned to the PMs are assigned to different PMs, and when any PM abnormally terminates due to a runaway or the like and the standby GSs are switched to active Even in, the load is not concentrated on one PM.

Next, each of the active placement unit 251 and the standby placement unit 261 determines whether or not the placement of all GS has been completed (step A6), and in the case of a negative determination, the processes of step A3 and subsequent steps are repeated. If the determination is affirmative, the GS placement process ends.

In this way, by sequentially arranging the active GS and the corresponding standby GS, it is possible to prevent the active and standby GS from being redundantly allocated to the same PM as a whole,
In addition, each GS can be distributed and arranged in each PM, and the arrangement can be automated.

FIG. 8 shows a rearrangement state when an abnormality occurs in PM0. Since the processing by the current GS1, GS5 and GS9 included in PM0 is no longer possible, each standby becomes a new current and the handover processing is performed. In addition, new working
New standby GS1, GS5 and G compatible with GS1, GS5 and GS9
With S9 placed, GS4, G for standby included in PM0
S7 and GS10 are newly placed.

In the embodiment of the present invention described above, four PM
Although the case of arranging 12 GS was considered, the arrangement can be similarly automated for any number and combination.

〔The invention's effect〕

As described above, according to the present invention, the active server space and the standby server space are arranged in the processors that are deviated by the count value, and the count values are updated, so that a plurality of processors arranged in the same processor are arranged. It is possible to allocate the standby server space corresponding to the current server space to different processors, and it is possible to automate the allocation of the server space in consideration of load distribution, which is extremely useful in practice. .

[Brief description of drawings]

FIG. 1 is a block diagram of the principle of the server space arrangement method of the present invention, and FIG. 2 is a configuration diagram of an embodiment to which the server space arrangement method of the present invention is applied, FIGS. 3, 4, and 5 FIG. 6 is an operation explanatory diagram of one embodiment, FIG. 6 and FIG. 7 are explanatory diagrams of a specific example of arrangement in one example, and FIG. 8 is an explanatory diagram of a specific example of rearrangement in one example. In the figure, 111 is a server space designating unit, 121 is a current arranging unit, 131 is a count value holding unit, 141 is a standby arranging unit, 211 is a PM configuration information storage unit, 221 is a server startup information storage unit, and 231 is a PM determination unit. 241 is a server determination unit, 251 is an active placement unit, 261 is a standby placement unit, 271 is a counter, and 281 is a placement information storage unit.

Claims (1)

(57) [Claims] 1. In a server space arranging method in which a plurality of server spaces each having an active and a standby are distributed and arranged in a plurality of processors, the active and standby server spaces to be arranged are sequentially arranged. A server space designating means (111) for designating, an active allocating means (121) for sequentially allocating the active server space sequentially designated by the server space designating means (111) to the adjacent processors, A count value holding means (131) for holding a count value related to the arrangement of the standby server space, and when the active server space is arranged by the active arrangement means (121), the corresponding standby server space is For the processor where the active server space is placed,
Arrangement in the processor shifted by the count value held in the count value holding means (131), and arranging for standby (141) for updating the count value every time the processor in which the active server space is placed makes a cycle. ), And a configuration for arranging the server space.