Electrical computers and digital processing systems: memory – Storage accessing and control – Hierarchical memories
Reexamination Certificate
2001-08-07
2004-04-13
Padmanabhan, Mano (Department: 2188)
Electrical computers and digital processing systems: memory
Storage accessing and control
Hierarchical memories
C711S114000, C711S162000, C711S165000, C707S793000, C710S032000
Reexamination Certificate
active
06721851
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates generally to data protection and more particularly to a system and method for solving the problem of sector slipping in a Storage Area Network.
2. Description of the Prior Art
Recent developments in storage solutions have led to the increased utilization by enterprises of Storage Area Networks (SANs) to provide storage consolidation, reliability, availability, and flexibility. Factors driving these developments include the increase in the amount of on-line data, data protection requirements including efficient and reliable data back-up, and rapidly increasing disk bit densities.
As illustrated in
FIG. 1
, an Information Technology (“IT”) Organization generally designated
100
includes a SAN
110
coupled between storage devices
120
and servers
130
. A Local Area Network (“LAN”)
140
networks clients
150
to servers
130
. The SAN
110
is conventionally a high-speed network that allows the establishment of direct connections between storage devices
120
and servers
130
. In the illustrated IT Organization
100
, the SAN
110
is shared between servers
130
, and allows for the sharing of storage devices
120
between the servers
130
providing greater availability and reliability of storage.
Third party copy is a method of transferring data directly between storage devices
120
in a SAN
110
using a data mover
200
such as illustrated in FIG.
2
. Data mover
200
may be disposed within a storage router or another SAN network component (not shown) or within a storage device such as disk array
220
. The connection between the client or application server
210
,
230
and the data mover
200
is conventionally a channel protocol like Small Computer System Interface (“SCSI”) or fibre channel connected directly to the storage devices
220
or storage device controllers (e.g. RAID controllers).
Data mover
200
is capable of initiating and controlling data movement on the SAN
110
at the direction of commands issued by other devices on the SAN
110
. To initiate data transfer from a SAN source storage device, such as tape drive
240
, to a SAN destination storage device, such as disk array
220
, an application server such as server
210
, issues a copy command to data mover
200
. The application server
210
manages the control information for the data transfer while the data mover
200
performs the actual data transfer from device
240
to device
220
. The application server
210
conventionally has ownership of a file system or database that resides on the SAN destination storage device
220
.
As illustrated in
FIG. 2
, the storage devices
220
and
240
are coupled to the SAN
110
, the SAN
110
including the data mover
200
. Alternatively, and as illustrated in
FIG. 3
, the SAN source storage device, such as a tape drive
340
, may be directly coupled to the SAN
110
through data mover
300
. A proprietary system, such as illustrated in
FIG. 4
, includes a data mover
400
coupled between the source storage device
410
and the destination storage device
420
. While the data movers
200
,
300
, and
400
have been illustrated as independent devices, it will be appreciated by those skilled in the art that data movers may be functionally implemented in storage device controllers.
Storage devices are conventionally designed to provide data to servers using one of two methods, either block-level or file-level access. Applications are optimized for either type of I/O access and both types of I/O access are usually supported within a customer site. File-level I/O is typically associated with LAN-based access while block-level access is associated with SAN-based access.
To initiate third party copy data transfers in the SAN
110
, the client or application server
210
generally provides the data mover
200
(
FIG. 2
) with the addresses of the source and destination devices and a list of data extents that describe the destination location. In the case of a block-to-block data transfer, both source and destination extents are specified. The extents include the starting location of the data blocks and the number of blocks to be transferred.
For the purposes of the present specification, the destination device for the data movement is a block (disk) device on which a file system or database resides and the source of the data can be any block or stream device (a serial device, i.e., a tape drive).
Due to the capability of file systems and database management systems to reorganize or write to the data residing on the destination device asynchronously of the third party copy operation, there is considerable risk in moving data into a live file system or database. The potential error conditions that arise due to a reorganization of the destination device occur after an extent list initiated by a third party copy request has been generated and sent to the data mover
200
. The potential error conditions are referred to as sector slipping events and manifest themselves as two error states on the destination block storage device.
A first sector slipping error state involves a movement of data or allocated space from the destination extents to another physical location (volume reorganization). As illustrated in
FIG. 5
Volume A includes destination blocks
510
corresponding to destination extents that are to be written by a third party copy operation. Destination blocks
510
are shown as being initially located or allocated on Disk
1
500
. Some time after the list of data extents has been provided to data mover
200
, but before the third party copy operation has completed, an error is detected on Disk
1
500
which causes the volume manager to move all data from Disk
1
500
to Disk
2
530
.
Since the third party copy operation has not yet completed and the destination blocks
510
have moved, there exists the possibility that the destination blocks
510
moved from Disk
1
500
to Disk
2
530
will not reflect all the data intended to be copied that is being written by the third party copy. Furthermore, the copy manager that is doing the block copy has no way of knowing that the reorganization is taking place and continues to move blocks into the destination blocks
510
on Disk
1
500
rather than to blocks
540
on Disk
2
530
even though the volume has been moved.
A second sector slipping error state involves the overwriting of data following a volume reorganization. With reference to
FIG. 6
, destination blocks
600
, located on Disk
1
610
, are to be written by a third party copy operation. While the third party copy operation is in progress, the destination blocks may be concurrently written by application “A” data
620
. This situation occurs generally due to a reallocation of disk space by an operation such as a disk optimization. Since the copy operation continues to write data to destination blocks
600
, the data stored by application “A” may potentially be corrupted and unreliable.
What is needed is a system and method for solving the problem of sector slipping when writing data into a live environment.
SUMMARY OF THE INVENTION
The present invention includes a block protection scheme within the block storage array to prevent a third party copy operation from writing data into locations that have become invalid due to a sector slipping event. The block protection scheme includes stalling any write operation while awaiting the cancellation of the third party copy operation. After the cancellation of the third party copy operation the original write from the host is allowed to complete.
In another aspect of the invention, an algorithm provides a stable copy into a live file system or database using a third party copy operation. The algorithm detects any changes in the data allocation that are not detected by the block protection scheme.
These and other features of the invention, as well as additional objects, advantages, and other novel features of the invention, will become apparent to those skilled in the art upon reading the following detailed description and acc
Kivlin B. Noäl
Meyertons Hood Kivlin Kowert & Goetzel P.C.
Padmanabhan Mano
Veritas Operating Corporation
Vital Pierre M.
LandOfFree
System and method for preventing sector slipping in a... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System and method for preventing sector slipping in a..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and method for preventing sector slipping in a... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3202938