Electrical computers and digital processing systems: memory – Storage accessing and control – Hierarchical memories
Reexamination Certificate
1998-01-07
2004-04-20
Kim, Matthew (Department: 2186)
Electrical computers and digital processing systems: memory
Storage accessing and control
Hierarchical memories
C711S112000
Reexamination Certificate
active
06725331
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to data storage systems. More particularly, the present invention is directed to a method and apparatus for managing the dynamic assignment of resources in a data storage system.
DESCRIPTION OF THE RELATED ART
Storage systems including devices such as disk drives are used in many different types of computer or data processing systems to store data. Disk drives generally include one or more disks of a recording medium (e.g., a magnetic recording medium or an optical recording medium) on which information can be written for storage purposes, and from which stored information can be read. Large data storage systems commonly include on the order of one-hundred disk drives, with each disk drive including several disks. One such mass storage system is the SYMMETRIX line of disk arrays available from EMC Corporation of Hopkinton, Mass. The SYMMETRIX line of disk arrays is described in numerous publications from EMC Corporation, including the SYMMETRIX model 55XX product manual, P-N200-810-550, rev. F, February, 1996.
Typically, data in a mass data storage system is accessed from a host computer in units called “logical volumes,” with the host computer writing or reading data to the storage system using a logical volume address or “logical device volume number” (hereafter DV#). Each physical storage device (e.g., a disk drive) in a storage system may store a single logical volume. Alternatively, it is possible in many systems to configure each physical storage device to store two or more logical volumes. For example, each disk drive may be configured to have two logical volumes stored on it, in which case the system would be said to have a two-to-one logical-to-physical relationship.
Some mass data storage systems allow for one or more resources of the system to be dynamically assigned during operation of the system, resulting in a change from the way these resources were allocated at the time the system was initially configured. One example of this involves the dynamic assignment, during operation, of one or more of the system's physical devices (e.g., a disk drive or a portion thereof) to store a particular logical volume. To support such dynamic assignment, one or more physical devices may be configured so that they are not statically assigned to store a particular logical volume addressable by the host. Instead, these storage devices are configured to be reserved for special operations during which they may be dynamically assigned by the storage system to temporarily store a particular logical volume addressable by the host. This dynamic assignment may be of the entire physical device, or some portion thereof.
One example of the dynamic assignment of physical resources involves configuring certain storage devices to be reserved as “hot spares” that are available to be dynamically assigned to, for example, replace failing storage devices. When it is determined that the number of errors occurring on a particular storage device is excessive and that a hard failure is probable, the storage system can dynamically assign an available one of its hot spares to address the problem. First, the system dynamically assigns the hot spare device as an additional mirror of the logical volume stored on the failing device, so that accesses to the logical volume can be serviced by the spare. The system then dynamically copies the data from the failing device to the available spare without interrupting operation of the storage system. The failing device is then replaced. Thereafter, the system dynamically copies the data from the hot spare to the newly installed device. Once the data is copied to the replacement device, the storage system dynamically reassigns the hot spare device so that it no longer acts as a mirror for the logical volume that was stored on the failing device and is returned to the pool of available spares for use in addressing potential failures with other devices.
Applicant has discovered that a number of problems can arise in handling the dynamic assignment of resources in a storage system, particularly when failures occur in the components of the system that store the information regarding the dynamic assignments. Several specific examples of the ways in which problems can arise are discussed below. However, before discussing those examples, an explanation is provided of the architecture of an existing SYMMETRIX data storage system, and the manner in which the dynamic assignment of devices is handled in that system. It should be appreciated that the challenges associated with the dynamic assignment of devices are not peculiar to the SYMMETRIX architecture, and apply to all types of storage systems that support the dynamic assignment of resources. Thus, the description of the existing SYMMETRIX system is provided below merely for illustrative purposes.
FIG. 1
is a block diagram of an existing SYMMETRIX data storage system
1
coupled to a host data processor
2
. The data storage system includes a number of disk drives (e.g., drives
112
,
114
,
116
,
118
,
132
,
134
,
136
, and
138
) and a number of controllers. Some of the controllers
102
,
104
are referred to as host adapters, and others (
110
A,
110
B,
130
A and
130
B) are referred to as disk adapters or DA's. The host adapters and disk adapters operate together, along with a read-write memory
100
that is globally accessible to all of the host and disk adapters, to transfer data between the host data processor
2
and the disk drives. It should be appreciated that the data storage system typically will include many more host adapters, disk adapters and disk drives than are shown in FIG.
1
. It should further be appreciated that each host adapter, disk adapter, and disk drive typically has a resident processor (e.g., a microprocessor) and local memory that are used to control its operation. The host adapters
102
and
104
communicate with the host data processor
2
via a bus
3
.
Communication between the host adapters
102
and
104
, the disk adapters
110
A,
110
B,
130
A and
130
B, and the globally accessible memory
100
is accomplished over busses
106
and
108
. Each of the disk adapters is coupled to a subset of the disk drives (
112
,
114
,
116
,
118
,
132
,
134
,
136
, and
138
) in the system, and controls communication with the drives to which it is coupled in a manner discussed briefly below. The disk adapters communicate with their respective disk drives via one or more buses
120
,
122
,
124
,
126
,
140
,
142
,
144
, and
146
, which may be of any type. For example, the buses can be SCSI (Small Computer System Interface) buses.
Globally accessible memory
100
includes three sections. A first section stores a number of tables used by the host and disk adapters to control communication between the host processor
2
and the disk drives (e.g.,
112
,
114
). A second section serves as a data cache that stores read/write data blocks, i.e., the system illustrated in
FIG. 1
is a cached storage system. The third section is used as a service area for a service processor
148
.
The manner in which the host adapters
102
and
104
and disk adapters
110
A,
110
B,
130
A and
130
B operate to enable the host data processor
2
to read data from and write data to the disk drives in the cached system of
FIG. 1
will now briefly be described. The caching operations are performed by the host adapters and disk adapters in a manner that is transparent to the host data processor. A read operation typically causes one of the host adapters
102
,
104
to scan a directory in the data cache portion of the memory
100
for the requested data, and when the requested data is in the cache, the host adapter transfers the data from the cache to the host processor
2
. If the requested data is not in the cache, the disk adapters (e.g.,
110
A,
110
B) determine on which disk drive (e.g.,
112
,
114
) the data is stored, and transfer the data from the disk drive to the cache. One of the host adapters then transfers the requested da
Choi Woo H.
EMC Corporation
Kim Matthew
Wolf Greenfield & Sacks P.C.
LandOfFree
Method and apparatus for managing the dynamic assignment... 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 and apparatus for managing the dynamic assignment..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for managing the dynamic assignment... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3263390