JPS6360418B2 - - Google Patents

Abstract

PURPOSE:To improve the reliability of a volume in a single-lung state when trouble occurs in both systems by collecting trouble statistics information previously and automatically, and selecting and disconnecting a more stable volume on the basis of the trouble statistics information in case of both-system trouble occurrence. CONSTITUTION:A trouble monitoring part 5 gathers statistics of temporary input/output trouble occurrence with respect to individual physical volumes. A disconnected volume selection part 6 when informed of the trouble of the physical volumes of the whole system with respect to a logical from a multiple volume control part 2 refers to the statistics information stored in a trouble statistics information storage part 11 to select physical volumes except the physical volume having the lowest frequency of trouble occurrence as objects of disconnection. Consequently, a volume disconnecting part 3 updates the contents of a multiple volume control table 8 and disconnects the informed physical volumes from the logical volume.

Description

DETAILED DESCRIPTION OF THE INVENTION (A) Technical Field of the Invention The present invention relates to a multiplexed volume failure handling method, particularly in a system having a function of duplicating and managing storage information in an external storage device, for example. When a volume fails in both systems of an individual physical device, the volume to be separated is determined based on the failure statistics information, and the reliability of the duplicated volume is improved in a single-lung state after a failure occurs in both systems. This is related to the processing method.

(B) Prior Art and Problems In recent years, information processing devices have been increasingly used in sophisticated ways, and improving their reliability has become an important issue. In particular, when an input/output failure occurs in an external storage device such as a magnetic disk drive, the information held by the device that encountered the failure is
To ensure that nothing is lost, support for so-called multiplexed volume control functions has been proposed in operating systems for medium and large aircraft.

In this multiplexed volume method, one virtual logical volume is associated with, for example, two physical volumes, and this correspondence is managed internally.
By making the user aware of only the logical volume, and writing the same content to each of the two physical volumes in response to a write request,
The same information is duplicated and retained. In the event of a failure in one, the other's normal volume can be accessed and information can be restored immediately if necessary, or operation can be continued using only the normal physical volume. When managing such multiplexed volumes, it is important to control the contents of each physical volume that corresponds to one logical volume so that there is no inconsistency.

By the way, for example, in a double volume,
Assume that an input/output failure occurs on both systems simultaneously. Although such a double failure rarely occurs, if it does occur, one physical volume should be separated to prevent discrepancies in the contents of each duplicated volume. The other physical volume performs what is called one-lung operation. Conventionally, when determining the physical volume to be disconnected at this time, the physical volumes were given slave relationships in advance, and the slave volume was selected for disconnection in the event of a failure in both systems. However, in this case, since the main volume is not necessarily highly reliable, the more stable sub-volume may be separated. Therefore, there has been a problem in that the reliability of the dual volume in a single lung state may be significantly reduced.

It is possible that the operator selects the isolated volume after a failure occurs in both systems, but in general, the operator should not be directly aware of the occurrence of a temporary input/output failure that can be recovered by retrying. It is often not possible to
There is a problem in that it is difficult to select a stable volume, and if operator intervention is required, automatic disconnection processing cannot be performed.

(C) Object and structure of the invention The present invention aims to solve the above-mentioned problems by automatically collecting failure statistical information in advance, and when a failure occurs in both systems, a more unstable volume is selected based on the failure statistical information. The aim is to improve the reliability of the volume in a single lung condition after the occurrence of a bilateral system disorder. Therefore, the multiplexed volume fault handling method of the present invention makes one logical volume correspond to a plurality of physical volumes, and performs multiplexed volume control so that each physical volume corresponding to the same logical volume has the same information. In a system that performs
a failure statistics information storage unit that stores failure statistics information of each of the physical volumes; a failure monitoring unit that monitors failures of each of the physical volumes and manages the failure statistics information stored in the failure statistics information storage unit;
When multiple failures occur in the physical volume, the storage information in the failure statistics information storage unit is referred to, a physical volume other than the physical volume with the lowest failure frequency is selected, and the correspondence relationship between the physical volume and the corresponding logical volume is determined. Equipped with a volume disconnection section that can be disconnected, in the event of multiple failures,
The system is characterized in that the logical volume associated with the fault is operated in association with only the physical volume with the least frequency of fault occurrence. Embodiments will be described below with reference to the drawings.

(D) Embodiment of the invention FIG. 1 shows the configuration of an embodiment of the invention, and FIG.
FIG. 4 is a process explanatory diagram of the illustrated device monitoring unit.

In the figure, 1 is a logical volume access section, 2 is a multiplexed volume control section, 3 is a volume separation section, 4 is a device monitoring section, 5 is a fault monitoring section, 6 is a separation volume selection section, 7 is a console display, 8 is the multiplexed volume management table, 9
is a logical volume, 10-1 and 10-2 are physical volumes, and 11 is a failure statistical information storage unit.

The logical volume access unit 1 writes data to or reads data from the logical volume 9 using, for example, a user program, and executes data processing. When accessing logical volume 9, logical volume 9 is regarded as one real storage medium, and physical volumes 10-1, 1 are accessed.
Don't be conscious of 0-2 etc. The multiplexed volume control unit 2 uses the multiplexed volume management table 8 to
Logical volume 9 and physical volume 10-1,1
0-2, and processes access requests from the logical volume access unit 1. Logical volume 9 and physical volume 10-
The correspondence relationship between 1 and 10-2 can be changed by, for example, a system configuration command by the operator.
Can give instructions. When each physical volume associated with one physical volume becomes a failure for one access to a logical volume, the volume separation unit 3 removes all physical volumes other than the physical volume with the least frequency of failure occurrence from the logical volume. It is separated from the volume.

The device monitoring unit 4 manages input/output devices such as physical volumes and monitors processing results of input/output requests to the physical volumes. It has a fault monitoring section 5 and a separation volume selection section 6. The failure monitoring unit 5 manages the failure statistical information storage unit 11,
Record I/O failure statistics for each physical volume. Generally, even if an input/output failure occurs, if the input/output processing is successfully completed by retrying, the input/output processing is incorporated into the multiplexed volume configuration as is. Fault monitoring section 5
Regarding such temporary input/output failure,
Obtain statistics for each physical volume.

When the separation volume selection unit 6 is notified from the multiplexing volume control unit 2 that a failure has occurred in the system-wide physical volume for the logical volume, the separation volume selection unit 6 refers to the statistical information stored in the failure statistical information storage unit 11, The number of failures that have occurred up to that point is checked for all physical volumes of the logical volume that caused the failure, and physical volumes other than the physical volume with the least number of failures are selected as targets for separation. Then, the selected volume information is notified to the volume separation unit 3. Thereby, the volume separation unit 3 updates the multiplexed volume management table 8 and separates the notified physical volume from the logical volume. Furthermore, information on the separated physical volume is displayed on the console display 7 to notify the operator of the separation.

The multiplexed volume management table 8 holds information on the correspondence between the logical volume 9 and the physical volumes 10-1 and 10-2. In this example, logical volume "A" is associated with physical volume "X" and physical volume "Y". Therefore, physical volume "X" and physical volume "Y" are managed so that they always have the same information. When physical volume "X" and physical volume "Y" become faulty at the same time, the fault-prone volume is separated as described above.
Therefore, by leaving a more stable physical volume, reliability in one-lung conditions increases, especially when there are differences in failure frequency.

The failure statistical information storage unit 11 is provided, for example, in a main storage device or an external storage device other than an input/output device to which a physical volume is allocated, and stores and holds the number of failure occurrences for each physical volume.

The processing outline of the apparatus monitoring section 4 shown in FIG. 1 is, for example, as shown in FIG. First, the process 20 shown in FIG. 2 waits for a fault notification from the multiplexed volume control section 2 or the input/output management section (not shown). When a failure notification is received, according to process 21,
For example, when physical volumes are duplicated, it is determined whether there is a failure in both systems. If it is not a double-system failure, the statistical information in the failure statistical information storage unit 11 is updated in process 22.

If there is a failure in both systems, control is transferred to process 23, and it is determined whether all retries of input/output processing for both systems have failed. If the retry is successful, control is returned to process 22. If all retries have failed, the process 24 refers to the failure statistics information storage unit 11, selects a physical volume with a large number of failures, and the process 25 causes the volume disconnection unit to disconnect the selected physical volume. Notify 3.

(E) Effects of the Invention As explained above, according to the present invention, for example, in a dual volume system, when an input/output failure occurs in both systems, single-lung operation can be continued while leaving a more stable physical volume. , reliability can be improved.

[Brief explanation of drawings]

Figure 1 shows the configuration of one embodiment of the present invention, and Figure 2 shows the configuration of the first embodiment.
FIG. 4 is a process explanatory diagram of the illustrated device monitoring unit. In the figure, 1 is a logical volume access section, 2 is a multiplexed volume control section, 3 is a volume separation section, 4 is a device monitoring section, 5 is a fault monitoring section, 6 is a separation volume selection section, 7 is a console display, 8 is the multiplexed volume management table, 9
is a logical volume, 10-1 and 10-2 are physical volumes, and 11 is a failure statistical information storage unit.

Claims (1)

[Scope of Claims] 1. In a system that performs multiplex volume control such that one logical volume corresponds to a plurality of physical volumes and each physical volume corresponding to the same logical volume has the same information, a failure statistics information storage unit that stores failure statistics information of each physical volume; a failure monitoring unit that monitors failures of each of the physical volumes and manages the failure statistics information stored in the failure statistics information storage unit; When multiple failures occur, the storage information in the failure statistics information storage unit is referred to, and a physical volume other than the physical volume with the least frequency of failures is selected, and the correspondence relationship between the physical volume and the corresponding logical volume is separated. 1. A multiplexed volume failure handling method characterized in that when multiple failures occur, the logical volume associated with the failure is operated in association with only the physical volume with the least frequency of failure occurrence.