JPH0736761A - On-line copying processing method with high reliability for external memory device - Google Patents

Abstract

PURPOSE:To always hold data in a duplicated state even in file recovery processing and on-line garbage processing after a device fault is restored by attaching priority on general access and copy processing, and inserting the general access to the copying processing. CONSTITUTION:A new member logic device 22 is generated by securing an area of same size as that of a member logic device 11 of a system in current use in an external memory device 2 of a spare system, and all data are copied on the area. When a write request by the general access is issued until the new member logic device 22 is erased after copy is started, a master logic device 31 performs write processing by the general access even on the new member logic device 22. When the copying on the new member logic device 22 is completed, the member logic device 21 of a spare system is deleted once, and the area is generated according to the information of the member logic device 11 of the system in current use from the leading position of a null area 6 near to the one with smaller number, and the copying is performed, and when the copying is completed, a newly generated member logic device 22 is erased.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION The present invention relates to a system for use in exchange processing, transmission processing, information processing, etc., which has a processor, a main storage device and a multiplexed external storage device, and always stores a plurality of data in the external storage device. The present invention relates to an on-line reliable copy processing method for existing at a location.

[0002]

2. Description of the Related Art In the prior art, a device multiplexing function has been performed in which the same data is multiply stored in a plurality of external storage devices. This is a high-reliability technology that substitutes a spare device when a failure occurs in the external storage device during normal operation. When executing the online garbage collection process for rearrangement, a state of single operation was generated. In order to prevent the deterioration of reliability during the single operation, there is a method of further increasing the reliability by adding an external storage device, but this leads to an increase in the cost of the (hardware) device. Therefore, it is necessary to prepare an online highly reliable copy processing method that effectively uses the existing apparatus.

[0003]

SUMMARY OF THE INVENTION In order to further improve the reliability of data in an external storage device in a system that is operated without interruption, the present invention provides a device multiplexing function that is currently being implemented. It is necessary to provide an online high-reliability multiple processing function that always uses multiple data without using any existing external storage device without expanding the device and without delaying the processing of general access. is there.

[0004]

As shown in FIG. 1, the present invention comprises a plurality of members each having a processor 4 having a main storage device 5 and external storage devices 1 and 2, and corresponding to the active external storage device 1. In a system having a plurality of member logical units 21 corresponding to the logical unit 11 and the standby external storage unit 2 and a master logical unit 31 for controlling the respective member logical units 11, 21, the external storage units 1, 2 are controlled. Under the control of the master logical unit 3, the member logical unit 1
Online reliable copy processing that saves backup information in the free space 6 of the backup external storage device 2 when writing new data to the backup external storage device 2 during online garbage collection processing or file restoration processing for assembling 1, 21 I do.

A problem in implementing the online high-reliability copy processing method is that, in addition to the configuration at the time of operation, the newly secured backup member logical device for backup is also controlled by the master logical device. In a general access write process, writing is performed in three member logical devices. As shown in FIG. 2, in the conventional method of device multiplexing, parallel access (# 1) is possible when the external storage devices are different, but parallel processing (# 2) is performed for the same external storage device. Is not performed, one general access (write process) is processed more than during normal operation (# 3)
Takes time. Also, when restoring a failed device or performing online garbage collection, copying processing is mixed in addition to these general accesses, which affects the delay of the general access. Therefore, by delaying the general access by adding a priority to the general access and the copying process and performing a scheduling for inserting the general access during the copying process.

[0006]

According to the present invention, since the data can always be held in the duplicated state even during the file restoration process after the device failure recovery and the online garbage process, the high reliability of the external storage device can be secured.

[0007]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described in detail below with reference to the drawings. FIG. 1 is a block diagram of the configuration of an external storage device according to an embodiment of the present invention. The external storage device has an active system 1 and a standby system 2, each of which has a plurality of member logical devices 11 and 21. Processor 4, main memory 5
Irrespective of the system configuration, the master logic unit 3 for controlling the external storage device and the master logic unit 31 for controlling the member logic device control the command to the external storage device from the upper software. During normal operation, the active member logic unit 11,
The standby system member logic device 21 and the master logic device 31 for controlling the member logic device are configured as one set, and a plurality of them are created by the upper software.

In the online high-reliability copy processing method, when the standby system member logical unit 21 is disconnected from the control of the master logical unit 31 under a certain condition (for example, area packing or restoration after recovery of a failed unit), the standby system is previously set. An area having the same size as the current active member logical device 11 is secured in the free space 6 in the external storage device 2, and all the data in the current active memory 11 is copied to maintain data duplication.

3 to 8 show the execution sequence of the online garbage collection process using the highly reliable copy processing method. Note that, in FIGS. 3 to 8, the same elements corresponding to each other are designated by the same reference numerals. Further, FIG. 9 shows the flow of the entire processing from FIG. 3 to FIG. When online garbage collection is requested, Figure 3
When there is a free area between the member logical devices in the active external storage device 1 as shown in (1), the active member logical device 11 next to the free space becomes the target member logical device. The same data is stored in the active member logical unit 11 and the standby member logical unit 21. Next, as shown in FIG. 4, an area of the same size as the active member logical device 11 is secured in the spare external storage device 2, a new member logical device 22 is created, and all the data is copied ($ 1). If there is a write request by general access from the start of the copying ($ 1) to the erasing of the new member logical unit 22, the master logical unit 31 is set to the new member logical unit 22. Also performs write processing by general access. When the copying ($ 1 in FIG. 4) to the new member logical unit 22 in FIG. 5 is completed, the spare member logical unit 21 in FIG. 4 is once deleted, and a free area closer to the younger one (6 in FIG. 4). An area is created from the beginning of 6) in FIG. 5 according to the information of the active member logical unit (11 in FIG. 5), and copying ($ 2 in FIG. 5) is performed, and the copying ($ in FIG. 5) is performed.
When step 2) is completed, the newly created member logical unit (22 in FIG. 5) is erased.

Next, as shown in FIG. 6, the active and standby external storage devices 1 and 2 are switched, and a new member logical device 12 is created in the switched external storage device 1.
Copy processing from the currently used member logical unit 21 ($
Perform 3). When the copying ($ 3) to the new member logical unit is completed, the switched standby member logical unit 11 is once deleted, and as shown in FIG. 7, it is switched according to the information of the switched active member logical unit 21. Copy ($ 4) to the standby member logical unit 11. After the copying ($ 4) is completed, the new member logical unit 12 is erased. Thereafter, the remaining member logical units 11 and 21 shown in FIG. 8 are similarly packed. After the garbage collection is completed, it returns to normal operation.

As shown in FIG. 10, when a failure occurs in the entire active external storage device 1 during the online garbage processing, the online garbage collection copying process ($ 5) is interrupted and the device multiplexing function is used. Although general access is performed from the standby external storage device 2,
Since the data of the member logical unit 21 during the online garbage collection is not reliable, general access is performed from the new member logical unit 22 as a substitute for the target member logical unit 21. As shown in FIG. 11, when the failure of the active external storage device 1 is recovered, the active external storage device is restored ($ 6) according to the logical device information in the standby external storage device 2. At this time, the target member logical unit 11 is restored from the new member logical unit 22 ($ 7). Further, in the restoration processing of the external storage device 1 that has been restored from the failure, the restoration processing is performed while assembling so as not to create an empty area, as in the case of executing the online garbage collection.

FIG. 12 shows the overall flow of processing when a failure occurs. If the failed system is a spare system, a new member logical unit cannot be guaranteed, so online garbage collection is stopped and only restoration processing from the active system member logical unit to the standby system member logical unit after failure recovery is performed. If the failed system is the current system, it is necessary to suspend the online garbage collection and restart the online garbage collection after the failure recovery. The reason is that a new logical unit is used as a substitute, so the current online member logical unit and the standby member logical unit are returned to the normal operation state by the restarted online garbage collection processing, and the new member logical unit is reduced. It is necessary.

FIG. 13 shows a scheduling method for general access and copy processing. As a scheduling method for controlling the execution of the general access queue 7 and the copy processing queue 8 requested by the upper software, the number of general access waits 71 to 75 is counted by the counter 9.
Here, if there are less than three general access waits 71 to 75, general access 71 to 75 and copy processing 81 to 83 are alternately executed, and if there are three or more, only general access 71 to 75 are executed. An example is a scheduling method of performing a threshold control of 3.

In FIG. 13, since the general accesses 71 to 75 are waiting for five processes in the order of write request 71, read request 72, write request 73, write 74, and read request 75, the count of the counter 9 is 5. . The general accesses 71 to 73 are executed first until this count becomes smaller than the threshold value. When the execution of the second write request 73 is completed, the count of the counter 9 becomes 2, so the first copy processing 81 is executed. Thereafter, the write request 74 for general access, the second copy processing 83, the read request 75 for general access, and the third copy processing 83 are executed in this order. The threshold can be changed for each system.

[0015]

As described above, according to the present invention, in a system having an external storage device of a current system and a standby system, it is possible to always hold data in a duplicated state even during file restoration processing after device failure recovery or online garbage processing. , High reliability can be secured.

[Brief description of drawings]

FIG. 1 is a block diagram of a system according to an embodiment of the present invention.

FIG. 2 is a diagram showing a processing sequence of the system of FIG.

FIG. 3 is a block diagram showing a processing progress of the present invention.

FIG. 4 is a block diagram showing a processing progress of the present invention.

FIG. 5 is a block diagram showing a processing progress of the present invention.

FIG. 6 is a block diagram showing a processing progress of the present invention.

FIG. 7 is a block diagram showing a processing progress of the present invention.

FIG. 8 is a block diagram showing a processing progress of the present invention.

FIG. 9 is a flowchart showing the processing of the present invention.

FIG. 10 is a block diagram showing the processing of the present invention.

FIG. 11 is a block diagram showing the processing of the present invention.

FIG. 12 is a flowchart showing a process when a failure occurs according to the present invention.

FIG. 13 is a block diagram of a scheduling method according to an embodiment.

[Description of Reference Signs] 1 external storage device (member physical device) 2 external storage device (member physical device) 3 master physical device 4 processor 5 main storage device 6 free area 7 general access queue 8 copy queue 9 threshold control Counter 11 Member logical unit 12 Newly created member logical unit 21 Member logical unit 22 Newly created member logical unit 31 Master logical unit 71 General access request 72 General access request 73 General access request 74 General access request 75 General Access request 81 Copy processing request 82 Copy processing request 83 Copy processing request

Claims (1)

[Claims] 1. A system adopts an operating mode of an active system / spare system, each system has an external storage device, and the inside of the external storage device is divided into a plurality of logical devices, each of the active system / spare system. In the copy processing of the external storage device that relocates the area online while the data area is guaranteed to have the same data in the same logical device, the data is either the active system or the standby system. Before using only one location, reliability is secured by using only existing equipment and performing copy processing, which is backup of data, in areas other than areas where data normally exists. In order to prevent delay in response of the general access, the priority is added to the general access and the copying process, and the scheduling for inserting the general access during the copying process is performed. Confidence copy processing method.