Methods and system for detecting data loss in a hierarchic data storage system

• US 5,572,661 A
• Filed: 10/05/1994
• Issued: 11/05/1996
• Est. Priority Date: 10/05/1994
• Status: Expired due to Term

• Current Assignee: Hewlett-Packard Company (HP Inc.)
• Original Assignee: Hewlett-Packard Company (HP Inc.)
• Inventors: Jacobson, Michael B.
• Primary Examiner(s): Beausoliel, Jr., Robert W.
• Assistant Examiner(s): Elmore, Stephen C.
• Application Number:
  US08/319,331
• Time in Patent Office: 762 Days
• Field of Search: 395/575, 395/425, 395/600, 395/275, 395/427, 395/650, 395/400, 395/182.05, 395/183.01, 395/183.18, 395/401, 395/404, 395/405, 395/406, 395/413, 395/417 , 395/441, 395/444, 395/490, 395/492, 395/497.01, 395/497.02, 365/200, 365/201, 364/474.16, 364/474.19, 364/550, 364/580, 364/738, 371/40.1, 324/512
• US Class Current: 714/6.32
• CPC Class Codes: G06F 11/1084 Degraded mode, e.g. caused by single or multiple storage removals or disk failures
• CPC Class Codes: G06F 2003/0697 device management, e.g. handlers, drivers, I/O schedulers
• CPC Class Codes: G06F 3/0601 Interfaces specially adapted for storage systems
• CPC Class Codes: G11B 20/1883 Methods for assignment of alternate areas for defective areas
• CPC Class Codes: G11B 2220/20 Disc-shaped record carriers
• CPC Class Codes: G11C 29/88 with partially good memories

Abstract

Data loss is selectively tracked within a hierarchic data storage system having multiple storage disks that define a physical storage space. The physical storage space is mapped into a first virtual storage space which presents the physical storage space as multiple RAID areas that store data according to various RAID levels, such as RAID Level 1 (mirror redundancy) or RAID Level 5 (parity redundancy). Individual RAID areas extend across multiple storage disks and include a region of physical storage space on individual storage disks. The first virtual storage space is mapped into a second virtual storage space which presents the physical storage space as multiple virtual blocks. The RAID areas are monitored to detect data loss. Upon detecting data loss within a region of a RAID area, this first region is marked as invalid. Upon detecting data loss within a second region of the same RAID area, the data storage system will determine the extent of data loss. If the data loss is comparatively minor, the data storage system will identify the virtual block that is mapped onto the portion of the second region and mark it defective. The address of the virtual block is stored in a virtual data loss table to indicate that data stored in that block is unreliable. If the data loss in the second region is significant, or if the virtual data loss table becomes full, the data loss will cause the entire RAID area to be classified as "dead".

Claims

1. A method for monitoring data loss in a hierarchic data storage system having multiple storage disks that define a physical storage space;

the method comprising the following steps;

(a) mapping the physical storage space into a first virtual storage space which presents the physical storage space as multiple RAID areas that store data according to various RAID levels, individual RAID areas extending across multiple storage disks and comprising a region of physical storage space on individual storage disks, individual RAID areas being initially classified as an optimal RAID area in which data can be stored and retrieved with redundancy;

(b) mapping the first virtual storage space into a second virtual storage space which presents the physical storage space as multiple virtual blocks;

(c) monitoring the RAID areas to detect loss of stored data;

(d) upon detecting data loss within a first region of a RAID area, marking the first region as an invalid region and classifying the RAID area as a degraded RAID area in which data can be stored and retrieved in that RAID area, but without redundancy; and

(e) upon detecting data loss within a second region of the degraded RAID area, classifying the RAID area as a dead RAID area in which data cannot be recovered from the first and second regions of that dead RAID area.

2. A method according to claim 1 further comprising the step of continuing to use other RAID areas that have not been classified as dead.

3. A method according to claim 1 further comprising the step of reporting to a user information concerning data loss as a result of the invalid regions and dead RAID areas.

4. A method for monitoring data loss in a hierarchic data storage system having multiple storage disks that define a physical storage space;

the method comprising the following steps;

(a) mapping the physical storage space into a first virtual storage space which presents the physical storage space as multiple RAID areas that store data according to various RAID levels, individual RAID areas extending across multiple storage disks and comprising a region of physical storage space on individual storage disks, individual RAID areas being initially classified as an optimal RAID area in which data can be stored and retrieved with redundancy;

(b) mapping the first virtual storage space into a second virtual storage space which presents the physical storage space as multiple virtual blocks;

(c) monitoring the RAID areas to detect loss of stored data;

(d) upon detecting data loss within a first region of a RAID area, marking the first region as an invalid region and classifying the RAID area as a degraded RAID area in which data can be stored and retrieved in the RAID area, but without redundancy;

(e) upon detecting data loss within a portion of a second region of the degraded RAID area, identifying a virtual block that is mapped onto the portion of the second region;

(f) indicating that the identified virtual block is a defective virtual block; and

(g) continuing to use remaining portions of the second region and any remaining regions within the degraded RAID area to store data without classifying the RAID area as dead.

5. A method according to claim 4 further comprising the additional step of listing the defective virtual block in a virtual data loss table.

6. A method according to claim 5 further comprising the following additional steps:

continuing to list defective virtual blocks in the virtual data loss table until the virtual data loss table is full; and

thereafter, upon detecting data loss within another portion of the second region of the degraded RAID area, marking the second region as invalid and classifying the RAID area as a dead RAID area in which data cannot be recovered from the first and second regions of the dead RAID area.

7. A method according to claim 5 further comprising the following additional steps:

writing data to the defective virtual block; and

removing the defective virtual block from the virtual data loss table and indicating that the virtual block is no longer defective.

8. A method according to claim 4 wherein prior to the step of identifying a virtual block in step (e), the following additional steps are conducted in lieu of steps (e)-(g) upon detection of data loss within a portion of the second region of the degraded RAID area:

first determining an extent of data loss within the second region;

marking the second region invalid in an event that a threshold quantity of the second region fails; and

classifying the RAID area as a dead RAID area in which data cannot be recovered from the first and second regions of the dead RAID area.

9. A method according to claim 4 further comprising the step of reporting to a user information concerning data loss as a result of the invalid regions and dead RAID areas.

10. A method for monitoring data loss in a hierarchic data storage system having multiple storage disks that define a physical storage space;

the method comprising the following steps;

(a) mapping the physical storage space into a first virtual storage space which presents the physical storage space as multiple RAID areas that store data according to various RAID levels, individual RAID areas extending across multiple storage disks and comprising a region of physical storage space on individual storage disks;

(b) mapping the first virtual storage space into a second virtual storage space which presents the physical storage space as multiple virtual blocks;

(c) monitoring the RAID areas to detect loss of stored data;

(d) upon detecting data loss within a region of a RAID area, identifying a virtual block that is mapped onto a portion of the region;

(e) indicating that the identified virtual block is a defective virtual block; and

(f) continuing to use remaining portions of the region and remaining regions within the RAID area to store data.

11. A method according to claim 10 further comprising the additional step of listing the defective virtual block in a virtual data loss table.

12. A method according to claim 10 further comprising the following additional steps;

listing a virtual address range of the defective virtual block in a virtual data loss table;

checking addresses involved in read operations against the virtual address range in the virtual data loss table; and

returning an error signal when the address involved in a read operation overlaps the virtual address range.

13. A method according to claim 10 further comprising the following additional steps:

writing data to the defective virtual block; and

removing the defective virtual block from the virtual data loss table and indicating that the virtual block is no longer defective.

14. A method according to claim 10 wherein prior to the step of identifying a virtual block in step (d), the following additional steps are conducted in lieu of steps (d)-(f) upon detection of data loss within a region of a RAID area:

first determining an extent of data loss within the region; and

marking the region invalid in an event that a threshold quantity of the region fails.

15. A hierarchic data storage system comprising:

a disk array having a plurality of storage disks that define a physical storage space;

a disk array controller coupled to the disk array for writing data to and reading data from the storage disks;

a RAID management system operatively coupled to the disk array controller for mapping first and second virtual storage spaces onto the physical storage space of the storage disks;

the first virtual storage space presenting the physical storage space as multiple RAID areas that store data according to various RAID levels, individual RAID area extending across multiple storage disks and comprising a region of physical storage space on individual storage disks;

the second virtual storage space presenting the physical storage space as multiple virtual blocks; and

a data loss detector to selectively track data loss that occurs within individual RAID areas, the data loss detector including a region sync table which contains multiple indicators that represent whether individual regions are alternately (1) valid and have not experienced any data loss or (2) invalid and have experience some data loss.

16. A hierarchic data storage system comprising:

a disk array having a plurality of storage disks that define a physical storage space;

a disk array controller coupled to the disk array for writing data to and reading data from the storage disks;

a RAID management system operatively coupled to the disk array controller for mapping first and second virtual storage spaces onto the physical storage space of the storage disks;

the first virtual storage space presenting the physical storage space as multiple RAID areas that store data according to various RAID levels, individual RAID areas extending across multiple storage disks and comprising a region of physical storage space on individual storage disks;

the second virtual storage space presenting the physical storage space as multiple virtual blocks;

a data loss detector to selectively track data loss that occurs within individual RAID areas, the data loss detector being configured to detect data loss within a region of a RAID area and identify in the second virtual storage space a defective virtual block that is mapped onto a portion of the region; and

the data loss detector including a virtual data loss table to store a virtual address of the defective virtual block.

17. A hierarchic data storage system comprising:

a disk array having a plurality of storage disks that define a physical storage space;

a disk array controller coupled to the disk array for writing data to and reading data from the storage disks;

a RAID management system operatively coupled to the disk array controller for mapping first and second virtual storage spaces onto the physical storage space of the storage disks;

the first virtual storage space presenting the physical storage space as multiple RAID areas that store data according to various RAID levels, individual RAID areas extending across multiple storage disks and comprising a region of physical storage space on individual storage disks;

the second virtual storage space presenting the physical storage space as multiple virtual blocks;

a data loss detector to selectively track data loss that occurs within individual RAID areas, the data loss detector including an active write store that maintains a list of all pending active write operations in the data storage system, a write operation being removed from the active write store upon completion;

the disk array controller attempts to write data to the defective virtual block; and

the data loss detector removes the defective virtual block from the virtual data loss table unless the data was lost while the write operation was pending and listed in the active write store.

18. A hierarchic data storage system comprising:

a disk array having a plurality of storage disks that define a physical storage space;

a disk array controller coupled to the disk array for writing data to and reading data from the storage disks;

a RAID management system operatively coupled to the disk array controller for mapping first and second virtual storage spaces onto the physical storage space of the storage disks;

the first virtual storage space presenting the physical storage space as multiple RAID areas that store data according to various RAID levels, individual RAID areas extending across multiple storage disks and comprising a region of physical storage space on individual storage disks;

the second virtual storage space presenting the physical storage space as multiple virtual blocks;

a data loss detector to selectively track data loss that occurs within individual RAID areas, the data loss detector being configured to detect data loss within a region of a RAID area;

the data loss detector including a region sync table which contains multiple sync bits for corresponding regions, individual sync bits having a first binary value to indicate that a corresponding region is a valid region and has not experienced any data loss and a second binary value to indicate that a corresponding region is an invalid region and has experienced some data loss;

whereupon detection of data loss within a first region of a RAID area, a sync bit associated with the first region is set to the second binary value indicating that the first region is an invalid region and the disk array controller avoids writing data to that invalid region;

the data loss detector being further configured to identify in the second virtual storage space a defective virtual block that is mapped onto a portion of a region;

the data loss detector including a virtual data loss table to store a virtual address of the defective virtual block; and

whereupon detection of data loss within a second region of a RAID area that already contains an invalid region, the address of the defective virtual block is added to the virtual data loss table and the disk array controller continues to access data in remaining portions of the second region and remaining regions within the same RAID area.

19. A hierarchic data storage system according to claim 18 wherein:

the disk array controller includes non-volatile RAM; and

the region sync table and the virtual data loss table are resident in the non-volatile RAM.