Electrical computers and digital processing systems: memory – Storage accessing and control – Control technique
Reexamination Certificate
2001-10-31
2004-08-31
Elmore, Reba I. (Department: 2187)
Electrical computers and digital processing systems: memory
Storage accessing and control
Control technique
C711S154000
Reexamination Certificate
active
06785791
ABSTRACT:
CROSS-REFERENCE TO RELATED APPLICATIONS
This application is related to, and claims the benefit of priority to, Japanese Patent Application number 2001-097344, filed Mar. 29, 2001 in Japan, the contents of which are incorporated herein by reference.
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a processing method for copying between data regions and a memory system for copying between logical disk and other data regions. More particularly, the present invention relates to a processing method for copying between data regions and a memory system making immediate access to a data region possible in response to an instruction to copy a data region provided in a main center to a data region provided in a supplemental center in a remote system, in which the main center and the supplemental center are disposed to communicate with a remote location.
2. Description of the Related Art
Copying of a logical disk and other data region units is executed. For example, a logical disk is an apparent disk device constructed without relation to the physical restrictions on physical disks. Accordingly, a logical disk may be constructed physically from a plurality of physical disk devices, and a plurality of logical disks may be constructed on one physical disk as well. The form that is used is determined by content and performance parameters.
Such a representative system is called a RAID (for example, RAID 0, 1, 0+1, 3 or 5) rank or a RAID group. RAID generally refers to a Redundant Arrays of Independent Disks.
Copying between physical disks in a remote system is mainly used as backup for saving data, but it is also used for various other purposes. For example, there are cases where tests are run without interrupting work. In such a case, the logical disk used in the work is copied, and the copy destination logical disk is used by the test system. Even if damage should occur to the data in the test system using the copy destination logical disk, there is thereby no effect on the work using the copy source logical disk.
Alternatively, a logical disk used in online work is copied, and the copy destination logical disk is used for batch processing. The performance of the online work using the source logical disk is thereby not affected by the batch processing.
In this manner, the number of instances where logical disk-to-disk copying is required is increasing. During the copying of the logical disks, work must be stopped. For example, it takes around 2 hours to copy 9 gigabytes (or GB). In recent environments with a 24-hour continuous operation, however, it is desired to come close to zero for the work stop time by coming close to zero for the copying time.
FIGS. 17
to
20
are explanatory diagrams of the Prior Art. FIG.
17
and
FIG. 18
are explanatory diagrams of conventional separate dual volume systems.
FIG. 19
is an explanatory diagram of a conventional log structure system.
FIG. 20
is an explanatory diagram of a conventional concurrent copy system.
The methods below are known as logical disk data copying methods provided in conventional RAID systems.
(1) Separate Dual Volume System
In a system comprising a RAID device
91
and a host
90
as indicated in
FIG. 17
, when the host
90
makes an instruction to copy as indicated in
FIG. 18
, preparation (definition) of a copy destination volume (logical disk)
94
is executed. Next, copying to a secondary (copy destination logical disk)
94
from a primary (copy source logical disk)
93
begins (a dual state is created) in response to a dual creation start instruction. When the copying is complete (the dual status is complete), the creation of a pair is complete. When this dual status ends, an update processing request from the host
90
is reflected in the primary (copy source logical disk)
93
and the secondary (copy destination logical disk)
94
. The host
90
issues a dual pair separation instruction, thereby completing the copying of the logical volume in the state at that point. As indicated in
FIG. 17
, separate hosts (applications)
90
and
95
can use the primary (copy source logical disk)
93
and the secondary (copy destination logical disk)
94
.
As indicated in
FIG. 18
, the primary (copy source logical disk)
93
can be accessed even during the copy interval and the secondary (copy destination logical disk)
94
cannot be accessed during the copy interval.
(2) Log Structured File
As indicated in
FIG. 19
, there are logical disks A, B and C defined by the host. Each data block is stored on one physical disk
96
. Each logical disk A, B and C is defined with data pointers for the physical disk
96
. Accordingly, copying from the logical disk A to the logical disk B is executed by setting the pointer for the logical disk A to the pointer of the logical disk B.
Data B
2
updated on the logical disk B is stored on the physical disk
96
separately from the original data. By updating the pointer for the logical disk B to the pointer for the updated data B
2
, the logical disk B can be updated without modifying original data A
1
to A
6
on the logical disk A.
This method does not accompany copying of actual data, and by setting pointers, the copy operation completes immediately.
(3) Concurrent Copy
As indicated in
FIG. 20
, during the copying from a copy source disk
93
to a copy destination disk
94
, a host
90
reads the copy target data on a copy source disk
93
(A), and writes to a copy destination disk
94
using a data transfer function. The bitmap of the address where copying is complete is set to copy complete.
During copying, if there is an update request to an un-copied region of the copy source disk
93
, that data on the copy source disk
93
is withdrawn to a side file
92
(
1
). The side file is set to the bitmap address portion during the withdrawal (
2
). Then, the copy source disk
93
is updated (
3
). While copying the data withdrawn to the side file
92
, copy target data is read from the side file
92
, and this is written to the copy destination disk
94
using a data transfer function (
4
). With this method, references/updates to the copy source disk
93
from the host are possible during copying.
Nevertheless, there are the following problems with the Prior Art, solved by the present invention.
(1) It is not possible to access the copy destination disk during the copy interval in the dual system in FIG.
17
and FIG.
18
. Because of this, it is necessary to wait for the copying to complete to access to the copy destination disk. For example, it takes around 2 hours to copy 9 GB, so a waiting time of 2 hours occurs. Because of this, a scheduled operation is required.
(2) With the log structure system in
FIG. 19
, access is possible to the copy source logic disk and the copy destination logical in response to a copy instruction, but because the actual data is not copied, the advantage of a dual system with the data copied onto a separate system is not gained. More specifically, damage to the copy destination logical disk is also damage to the copy source logical disk, so if damage occurs to the copy destination physical disk, the copy source logical disc cannot be used either. Also, access to the copy destination physical disk is access to the copy source physical disk, so high speed performance through a distribution to separate disks is not obtained.
(3) With the concurrent copy system in
FIG. 20
, access to the copy source logical disk is possible during copying, but it is necessary to wait until copying completes to access the copy destination logical disk.
A similar problem is that to allow access from a host device, logical disk formatting was required beforehand, and until the formatting ended, access could not be made to the logical disk. The time required for the logical formatting differs according to the size of the logical volume, but it takes roughly tens of minutes to a number of hours, and during this period, data cannot be processed. With the current state of things, the time required for operations such as new installation of storag
Daikokuya Hidejiro
Ito Mikio
Onda Yasuhiro
Watanabe Futoshi
Elmore Reba I.
Fujitsu Limited
Staas & Halsey , LLP
LandOfFree
Method for accessing data being copied between data regions... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method for accessing data being copied between data regions..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for accessing data being copied between data regions... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3342514