JP2772052B2 - Resource information inheritance processing method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Overview] By transferring the information set in the nonvolatile shared memory from the active process to the standby process, when the active process fails, the standby process replaces the active process.
A resource information takeover processing method for fault-tolerant computer systems that continue to operate in order to maintain the performance of the current process during normal operation and improve switching performance. The resource information to be taken over to the standby process is determined according to the nature of the resource information change frequency.
The resource information of the first group is classified into a group and a second group, and the resource information of the first group is collectively taken over by a standby process when the active process crashes. so,
Alternatively, when the amount of change information reaches a predetermined amount, a process of intermittently taking over to a standby process is performed, and both batch takeover and intermittent takeover are used.

[Industrial applications]

According to the present invention, for example, in a computer system in which a standby process is replaced by an active process and a job is continuously operated when an active process fails, information set in a non-volatile shared memory is inherited from an active process to a standby process. The present invention relates to a resource information takeover processing method for fault tolerance.

[Conventional technology]

In a computer system, a method of duplicating a control process is effective in realizing fault tolerance. One is a control process that operates for normal operation, and the other is a control process that is on standby to continue operations when the control process for normal operation is stopped due to some failure. .

The former control process is called an active process, and the latter control process is called a standby process.

When switching the processing subject from the active process to the standby process due to a failure, it is necessary to take over various types of resource information. In order to speed up the transfer of the resource information, it is effective to perform the transfer between the active process and the standby process via a non-volatile shared memory such as a semiconductor external storage device that can be commonly accessed. .

The following two schemes can be considered as a method of taking over the resource information via the nonvolatile shared memory.

One is a method in which resource information is taken over at the same time when the active process crashes.

The other is a method of successively taking over resource information during operation of the active process.

[Problems to be solved by the invention]

By taking over resource information between the active process and the standby process via the non-volatile shared memory,
Switching can be accelerated to some extent,
As the amount of resource information increases, switching performance is affected.

In particular, in the case where the resource information is taken over at the same time when the active process crashes, it takes time to transfer and set all the resource information at the time of the crash, so that the switching performance is deteriorated.

On the other hand, if the resource information is successively taken over during the operation of the active process, an extra processing step is required every time the active process updates the resource information. There is a problem that causes trouble.

SUMMARY OF THE INVENTION It is an object of the present invention to provide means for maintaining the performance of a current process during normal operation and improving switching performance.

[Means for solving the problem]

 FIG. 1 is a diagram illustrating the principle of the present invention.

In FIG. 1, reference numeral 10 denotes a nonvolatile shared memory, 11 denotes a working process that operates for normal operation, and 12 denotes a working process 11.
In the event of a failure, a standby process waiting for continuous operation of the business, 13 indicates intermittent takeover processing, and 14 indicates batch takeover processing. A1, A2, B1 to B3 represent resource information to be taken over.

In the present invention, the resource information taken over from the active process 11 to the standby process 12 is classified into two groups according to the frequency of change. The first group is a group of resource information whose change frequency is relatively high. The second group is a group with a small change frequency. Here, the resource information means information that needs to be taken over for fault tolerance.

As shown in FIG. 1A, when the active process 11 changes resource information during normal operation, the active process 11 writes the changed resource information to the nonvolatile shared memory 10. Of these pieces of resource information, the first group takeover information A1 and A2 are stored in the nonvolatile shared memory 10 as they are.

With respect to the handover information B1 to B3 of the second group, a process of intermittently handing over the standby process 12 is performed at a predetermined time interval or when the amount of change information reaches a predetermined amount. In FIG. 1 (a), since the takeover information B1 and B2 have reached the predetermined values, the standby process 12 takes in the information in its own process.

If the active process 11 crashes due to a failure, the standby process 12 needs to take over all the resource inheritance information so that the work performed by the active process 11 can be continued by the standby process 12. Therefore, when the standby process 12 receives a crash notification of the active process 11,
As shown in FIG. 1 (b), a process for taking over the takeover information A1, A2, B3 that has not been completed during normal operation. The takeover information A1 and A2 of the first group are taken over collectively. As for the handover information of the second group, only the remaining information that has not been handed over in the intermittent handover process 13 is handed over.

[Action]

Takeover information A1, A2 of the first group that changes frequently
As for (2), when the active process 11 crashes, as shown in FIG. Therefore, the overhead for taking over the active process 11 during normal operation can be reduced.

Inheritance information B1 to B3 of the second group that changes less frequently
As shown in FIG. 1A, during the normal operation, the intermittent takeover process 13 is performed from the active process 11 to the standby process 12. Therefore, the current process 11
In the event of a crash, information that has already undergone intermittent takeover processing 13 can be skipped, so that the processing time required for takeover can be reduced by that much, and switching performance can be improved. Can be.

As described above, the batch takeover processing 14 and the intermittent takeover processing 13 are used in consideration of the change frequency of the resource information and the like, so that overhead during normal operation and crash can be efficiently reduced.

〔Example〕

FIG. 2 is an example of an application system of the present invention, FIG. 3 is an example of a data structure of takeover information according to an embodiment of the present invention, FIG.
FIG. 5 is a processing flow of a working process in a normal state according to an embodiment of the present invention, FIG. 5 is a processing flow of a standby process in a normal state according to an embodiment of the present invention, and FIG. 6 is an embodiment of the present invention. 2 shows a processing flow of a standby process according to the first embodiment.

The present invention can be applied to, for example, a complex system as shown in FIG. In FIG. 2, those having the same reference numerals as in FIG. 1 correspond to those shown in FIG. 1, 20 is a nonvolatile shared memory for collective takeover, 21 is a nonvolatile shared memory for intermittent takeover, 22-1, 22- 2 is a processor module (PM) equipped with a CPU, memory, etc., 23 is a handover information writing processing unit, 24 is an intermittent handover determination unit, 25 is an intermittent handover instruction unit, 26 is a batch handover unit at crash, and 27 is an intermittent handover unit. The takeover processing unit 28 represents a bus.

The processor modules 22-1 and 22-2 are physical units that can exclusively use the CPU and the memory. The bus 28 is used between each processor module and a non-volatile shared memory.
This is a data transfer path connecting between 10 and the processor module.

In the non-volatile shared memory 10, areas of a non-volatile shared memory for batch taking over in which batch taking over information is stored and a non-volatile shared memory for intermittent taking over in which intermittent taking over information is stored are provided.

In the example of FIG. 2, the active process 11 operates in the processor module 22-1, and the standby process 12 is arranged in the processor module 22-2 and stands by.

Inherited information write processing unit in active process 11
23 determines the type of the resource information in response to the write request of the takeover information due to the change of the resource information.
In the second group, a process of writing the data into the non-volatile shared memory 21 for intermittent handover is performed.

Further, the intermittent takeover determination unit 24 determines whether or not the intermittent takeover processing is necessary at present. When the change amount exceeds a predetermined amount and the takeover is required, the intermittent takeover instruction unit 25 determines whether or not the intermittent takeover processing is necessary. , And issues an intermittent takeover instruction to the standby process 12.

The standby process 12 carries out resource information takeover processing in response to a crash notification of the active process 11 from a failure detection unit (not shown) or an instruction from the intermittent takeover instruction unit 25. In response to the crash notification, the crash-time batch takeover processing unit 26 performs the process of taking over the resource information as shown in FIG. In response to the intermittent takeover instruction, the intermittent takeover processing unit 27 refers to the intermittent takeover non-volatile shared memory 21 and performs a resource information takeover process as shown in FIG.

The data structure of the batch takeover information stored in the batch takeover nonvolatile shared memory 20 is as shown in FIG. * 1 is a busy flag indicating whether the information is valid, and * 2 is a storage area of the information type.

The intermittent handover information stored in the intermittent handover nonvolatile shared memory 21 is managed by intermittent handover entries as shown in FIG. * 1 is a busy flag, and * 2 is a storage area of the information type. For each entry, one takeover information is registered.

Next, specific processing contents of the active process 11 and the standby process 12 according to one embodiment of the present invention will be described.

During normal operation, the active process 11 performs the processing shown in FIG. 4 in response to a request to write resource information to the nonvolatile shared memory 10. In the following description corresponds to the processing shown in FIG.

Non-volatile memory 20 for batch transfer shown in FIG.
It is determined whether or not the request is a write request to the host. That is,
It is determined whether the group is the first group (collective) or the second group (intermittent).

 In the case of intermittent handover, the processing shifts to processing.

In the case of batch transfer, a transfer information storage area is acquired from the batch transfer nonvolatile shared memory 20.

Turn on the busy flag in the handover information storage area, set the information type, and write the handover information requested to be written to the handover information storage area. Then, the process ends.

In the case of resource information subject to intermittent handover, a handover information storage area is acquired from the intermittent handover nonvolatile shared memory 21.

 Write the passing information to the passing information storage area.

Make the intermittent handover entry in use and enter the handover information storage area. Then, the busy flag in the handover information storage area is turned ON.

It is determined whether the number of intermittent takeover entries has reached a predetermined amount. If a certain amount is reached,
Execute the process.

If the fixed amount has not been reached, time monitoring by a timer is started as necessary. Thereafter, when a timeout is received, the processing is executed.

An instruction is issued to the standby process 12 to perform intermittent takeover.

Upon receiving the intermittent information takeover instruction from the active process 11 during normal operation, the standby process 12 executes the processing shown in FIG.

The takeover information is read from the takeover information storage area entered in the intermittent takeover entry as shown in FIG.

Based on the information type in the handover information storage area, the handover information is determined and reflected in the standby process 12. Then wait for the next instruction.

When the standby process 12 receives the crash notification of the active process 11, the standby process 12 executes the processing shown in FIG.

The intermittent handover entry is read from the intermittent handover non-volatile shared memory 21 and it is checked whether or not there is an entry of the handover information registered therein.

Whether or not there is an entry is indicated by the in-use flag. If there is an entry, that is, if the busy flag is ON, the process is executed, and if the busy flag is OFF, the process is skipped.

The information is read from the takeover information storage area entered in the intermittent takeover entry and reflected in the standby process 12.

The contents of the takeover information storage area are read from the non-volatile shared memory for batch takeover 20.

Determine whether the busy flag in the handover information storage area is ON. If ON, execute the following processing.

Apply the batch transfer information to the standby process 12.
When all of the above takeover processes have been completed, the work is resumed with the standby process 12 as the new active.

〔The invention's effect〕

As described above, according to the present invention, it is possible to maintain the performance in the normal operation of the active process and to improve the switching performance when the active process crashes.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating the principle of the present invention, FIG. 2 is an example of an application system of the present invention, FIG. 3 is an example of a data structure of takeover information according to an embodiment of the present invention, and FIG. FIG. 5 is a processing flow of a standby process in a normal state according to an embodiment of the present invention, and FIG. 6 is a processing flow of a standby process in a normal state according to an embodiment of the present invention. 4 shows a processing flow. In the figure, 10 is a nonvolatile shared memory, 11 is a working process, 12 is a standby process, 13 is an intermittent takeover process, 14 is a batch takeover process, and A1, A2, and B1 to B3 represent takeover information.

Continued on the front page (58) Fields surveyed (Int.Cl. ⁶ , DB name) G06F 11/16-11 / 20,12 / 16,15 / 16 G05B 9/03 H04M 3/22 H04L 11/20

Claims (1)

(57) [Claims] A working process (11) and a standby process (12).
By transferring the resource information necessary for taking over the business from the active process to the standby process, when the active process fails, the standby process replaces the active process and the resource in the computer system that continues to operate the business. In the information transfer processing method, the resource information transferred from the active process to the standby process is
A process in which the resource information of the first group is classified into a first group and a second group according to the characteristics taking into account at least the frequency of change of the resource information. The resource information of the second group is intermittently handed over to the standby process at a predetermined time interval or when the amount of change information reaches the predetermined amount. A resource information takeover processing method characterized by using together with takeover (13).