摘要:
Described is a distributed copying technique that may be used in migrating large amounts of data from one or more source devices to one or more destination devices. The data source is divided into partitions. As Fibre Channel adapters (FAs) become available, each of the FAs may copy a partition of the data. In connection with specifying paths used for the distributed copying technique, a preferred path selection (source port-target port mapping) may be made by executing code in a controlling data storage system to perform discovery processing. The preferred path selection is used for the duration of the distributed copying unless the preferred path is unable to transmit data. A target port of the preferred path may be randomly selected from all accessible target ports, and/or in accordance with a specified portion of the target ports. Preferred paths may also be specified using an API (application programming interface).
摘要:
Described are techniques for use with online data copying or migration. Data is pushed from a source to a target device while one or both of the source and target devices are available for use by one or more host systems. The host systems may perform I/O operations to the source and/or target devices while the data migration is ongoing if a copy session is defined on the data storage system(s) of the source and/or target devices. A copying operation performs data migration from the source to the target device. A migration vector is used to indicate which portions of the source have been migrated. A change vector may be associated with each of the source and/or target devices to indicate portions that have been modified in order to perform a differential copy operation.
摘要:
Described are techniques for use with an error handling policy for a data storage system. Error handling criteria may be specified for controlling behavior of the data storage system upon the occurrence of an internal processing error occurring when performing an ancillary task associated with a data operation received by the data storage system. The error handling criteria may include a threshold counter value, and one or more of: a device, a specific device location or address, and a host. An error may be conditionally returned to the host upon the occurrence of an internal processing error in accordance with the error handling criteria.
摘要:
Described is a distributed copying technique that may be used in migrating large amounts of data from one or more source devices to one or more destination devices. The data source is divided into partitions. As Fibre Channel adapters (FAs) become available, each of the FAs may copy a partition of the data. In connection with specifying paths used for the distributed copying technique, a preferred path selection (source port-target port mapping) may be made by executing code in a controlling data storage system to perform discovery processing. The preferred path selection is used for the duration of the distributed copying unless the preferred path is unable to transmit data. A target port of the preferred path may be randomly selected from all accessible target ports, and/or in accordance with a specified portion of the target ports. Preferred paths may also be specified using an API (application programming interface).
摘要:
Described is a distributed copying technique that may be used in migrating large amounts of data from one or more source devices to one or more destination devices. The data source is divided into partitions. As Fibre Channel adapters (FAs) become available, each of the FAs may copy a partition of the data. A copy indicator value indicates the last partition delegated to an FA for copying. The last FA to complete copying is responsible for ensuring completed copying of any other data partitions delegated to an FA which has died during migration of the other data partitions. In the event that the last FA dies prior to completing its processing, additional steps are performed by other remaining FAs to copy all data partitions not known to be successfully migrated.
摘要:
Described are techniques for allocating resources. Two consumer processes using two resources are provided. A first threshold of a first of the resources for a first of said two consumer processes is provided. A second threshold of a second of said resources for said first consumer process is provided. Consumption information is received regarding actual consumption of the two resources by the first consumer process. Execution of the first consumer process is scheduled in accordance with the first and the second thresholds and the consumption information.
摘要:
Two data operations, such as write operations, may be processed at a same time in which the two write operations operate on a same address range span. A first of the write operations may write to the first track and the last track in the span but not tracks therebetween. When processing the first write operation, a lock for the first track is obtained and the data is written to the first track data. A lock for the last track is obtained and it is determined whether all tracks between the first and last tracks are unlocked. If so, data is written to the last track. If not, steps of releasing and then reacquiring the lock for the last track and determining whether intervening tracks are locked are repeated until the intervening tracks are not locked. The last track is then written.
摘要:
The invention relates to a method and apparatus for reordering, at a disk drive controller, a received sequence of read and write commands. The reordering can be system driven so that the system, when a command stalls, examines the command queue for another command to execute. Alternatively, the command queue itself can be reordered depending upon the expected time it takes for various commands to execute. Reordered commands can be flagged so that they cannot be reordered again. The resulting increase in throughput and efficiency occurs without reference to the particular disk drive or disk drive array configuration to which the controller is connected.
摘要:
A method and apparatus, for a system having a plurality of host computers and a memory system having a disk controller with at least one mass storage device having a plurality of logical volumes, provide for write protecting the logical volumes at the controller port level. The method features changing the write protect status of one or more logical volumes, on a port level basis, in response to a first host request to change a write protect status; writing a write protect status message to a memory associated with the controller channel directors; polling, from each channel director, the memory for an unread write protect status messages; and preventing, at, for example, the channel director port level, a host connected to the channel director from writing to a protected logical volume which is write protected from that host or host port. The write protect status of that channel director, for that logical volume, remains in effect until the write protect condition for that volume has been reset. A lock mechanism must be obtained to make any changes in the logical volume write protect status.
摘要:
The invention relates to a disk drive controller system having redundant operations spread between at least two SCSI adaptors connected to a SCSI bus. At least one host computer is also connected to the SCSI bus. If one of the SCSI adaptors fails, the other SCSI adaptor connected to the bus, upon detecting the failure, takes over for the devices serviced by the failing SCSI adaptor. A special termination connection, in which one SCSI adaptor terminates the SCSI bus while the other SCSI adaptor does not, is provided for implementing the invention. A defined process enables the detecting SCSI adaptor to take over for the failed SCSI adaptor and for the failed SCSI adaptor to thereafter return to normal operation.