Disk array device including a memory with backup power...

Details Disk array device including a memory with backup power... Disk array device including a memory with backup power...

2001-02-15

2004-11-09

Beausoliel, Robert (Department: 2184)

Error detection/correction and fault detection/recovery

Data processing system error or fault handling

Reliability and availability

C714S014000, C711S114000

Reexamination Certificate

active

06816981

ABSTRACT:

CROSS REFERENCE TO RELATED APPLICATIONS
The contents of 2000-167484, filed Jun. 5, 2000 in Japan, are incorporated herein by reference.
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to disk array devices including multiple disk devices equipped with a restorative function used when the disk devices fail.
2. Description of the Related Art
In recent years, devices have been made available in which subsystem components are multiplexed and have a degree of redundancy. A localized trouble-shooting function and the redundant configuration of these devices increases the continuous usability of these devices and allows for the automatic restoration of data when a disk fails. The data redundancy method is categorized into 6 stages ranging from RAID 0 to RAID 5 (RAID is the Redundant Array of Inexpensive Disks).
FIG.
14
(
a
) shows a schematic diagram of a RAID 4 System
100
. As shown in FIG.
14
(
a
), RAID 4 employs a parity system for data restoration information. The RAID 4 System
100
shown in FIG.
14
(
a
) includes several data disks D
0
, D
2
, . . . storing the data allocated in multiple read/write units, a parity-generating unit P and a disk device DP which stores the parity.
In the RAID 4 System
100
shown in FIG.
14
(
a
), the data is allocated into multiple units designated by A
0
, A
2
, . . . . Generally, these data units A
0
, A
2
, . . . are fixed lengths. The allocated data A
0
, A
2
, . . . are distributed to and stored on the data disks D
0
, D
2
, . . . while the parity is stored on the dedicated disk DP. In the following description, the data that is allocated to different disks and stored as above is referred to as redundant identical group data or simply as redundant group data. The disk groups that this data is stored on are referred to as redundant group disks. The parity can also be referred to as redundant data.
When there is a problem with a disk, the data on the disk is regenerated from the remaining identical group data and parity (redundant data).
RAID 4 is capable of reading out multiple data simultaneously but cannot write multiple data at the same time. When updating data, the RAID 4 System
100
always reads the parity and the data before the update and writes after creating the update parity, which requires additional access. This is referred to as a write penalty.
FIG.
14
(
b
) shows a schematic diagram of a RAID 5 System
200
. Like the RAID 4 System
100
, the RAID 5 System
200
employs a parity system for data restoration information. The RAID 5 System
200
also includes multiple disks D
1
, D
2
, . . . for storing parity, and a parity-generating unit P.
In the RAID 5 System
200
, the data is divided into several groups as shown in FIG.
14
(
b
), including A
0
, A
1
, . . . B
0
, B
1
, . . . The groups of divided data are distributed to disks D
1
, D
2
, . . . respectively and stored therein. The parity PA, of the data A
0
, A
1
. . . and the parity PB of the data B
0
, B
1
. . . are distributed to the disks D
1
, D
2
, D
3
. . . and stored.
In the RAID 5 System
200
, as with the RAID 4 System
100
, when there is a problem with a disk, the data on the disk is regenerated from the aforementioned identical group data and parity (parity data).
RAID 5 is capable of reading and writing multiple disks simultaneously. When updating the data, there is the aforementioned write penalty. Also, while updating the parity, no read/write access is allowed to the disk.
FIG. 15
shows a diagram with an example of the write sequence in a disk array device to which the aforementioned RAID 4 System
100
or RAID 5 System
200
could be applied. The example of the write sequence shown in
FIG. 15
corresponds to the RAID 4 System
100
and is explained with reference to the RAID 4 System
100
.
As shown in
FIG. 15
, the RAID 4 System
100
includes a subsystem control module
101
, which includes subsystem control module internal memory
101
a
. Also shown in
FIG. 15
, the RAID 4 System
100
includes subsystem internal interface module
102
(hereafter, “interface” is abbreviated to I/F), device control module
103
, buffer
103
a
, device I/F module
104
, disk group
105
, data disks D
0
~D
2
, and redundant disk P storing redundant data.
Referring now to
FIG. 15
, OD (Old Data) is the data that is to be updated (referred to as old data below), OP (Old Parity) is the parity data to be updated (referred to as old parity below), ND (New Data) is the write data, NP (New Parity) is the write redundant data (referred to as new parity below) and IP is the interim parity data.
As shown in
FIG. 15
, the write operation in the disk array device
100
is carried out as explained herein below. (Items (a)~(g) in
FIG. 15
correspond, respectively, to items (a)~(g) below.)
(a) The write data ND
1
is transferred from the memory
101
a
of the subsystem control module
101
to the buffer
103
a
of the device control module
103
.
(b) The data OD
1
on the disk that is to be written to is read into the buffer
103
a.
(c) The redundant data OP of the redundant group of the data to be written is read into the buffer
103
a.
(d) The interim redundant data IP is generated by performing an exclusive “or” operation on OD
1
and OP.
(e) The new redundant data NP is generated by performing an exclusive “or” operation on ND
1
and IP.
(f) ND
1
is written to the disk
105
.
(g) NP is written to the disk
105
.
For this sequence, items (a)~(c) and (e)~(f) do not have to be performed in any strictly fixed order.
The following types of methods are possible for maintaining the reliability of the data when there has been a momentary interruption due to a power outage or other reason in systems that perform the above sort of write operation:
(1) Continuous subsystem operation by means of a battery back-up system for the entire device.
(2) Write data support based on non-volatile memory.
In (1) above, when the power supply supplied to the device is cut off, the data is secured by the continuous operation of a subsystem. However, in (1) above, a large-capacity battery is required to back up the entire subsystem and in actual installations, the percent that this occupies is extremely large.
In (2) above, the write data remains in memory which nearly always makes recovery possible by writing to the disk again when the power supply is turned back on. Also in (2) above, if the write was being carried out before power supply was cut off and the redundant data was being written, that RAID would be in degeneration mode (at least when one disk had failed). Then when the power supply was turned on again, when the RAID shifted into degeneration mode, the redundancy of that redundant group would be lost and it would not be possible (since it would not be performed properly) to restore the data on the broken disk or to write the data of that redundant group. This state is referred to as a “Write Hole”.
To correct this sort of problem, the redundant data stored in the memory
101
a
is managed constantly and the status of the write progress is written to the memory
101
a
. That progress status is used with the redundant data to perform the recovery. Avoiding the above state has also been considered, but the need to constantly transfer redundant data while writing led to a drop in performance.
SUMMARY OF THE INVENTION
The present invention solves the above-mentioned problems.
An object of the present invention is to provide a disk array device that maintains reliability of data without too great a loss of performance, even in degeneration mode, as well as when the power supply is turned back on after being turned off.
The present invention comprises a disk array device including a subsystem control module, disk, and a device control module. The subsystem control module comprises a memory backed up with a battery. The disks store data and/or parity. The device control module controls the disks. The device control module comprises a buffer storing redundant data, wherein when data is to be written to the disks, the disk array device allocates and wri

Affiliated with

Also associated with

Rating

Disk array device including a memory with backup power... does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Disk array device including a memory with backup power..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Disk array device including a memory with backup power... will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-3274813