摘要:
A system for asynchronously duplexing direct access storage device (DASD) data in a plurality of DASD subsystems has the advantage of decoupling the data duplexing operation from the DASD write I/O operation. This ensures the write does not incur unnecessary wait states in the subsystem. By establishing a sequence checkpoint at which time a set of information packets are grouped together and processed as a single sequence unit, this decoupling and independent operation takes place. Through this independence, data copying to a secondary location can take place without affecting the performance of the subsystems and also not affecting the corresponding integrity of the data that is being updated.
摘要:
A data copy validation technique compares primary data to a copy of that data by generating check codes of the data to be validated at both primary and secondary sites in a continuously running remote data shadowing system. The validation session runs concurrently with the data shadowing so as not to disrupt the data shadowing. Each data validation session is assigned a copy start time, for example, a time preceding a time of a consistency group, for defining when the check code for data at the primary should be compared to a check code of the data at the secondary. The check code of the data at the secondary site is taken at a time that is the equivalent logical point-in-time of that data at the primary site even though the respective data contents may be different in real time.
摘要:
A remote data shadowing system provides synchronous, storage based, real time disaster recovery wherein a secondary site is located remote from the primary site. An error recovery program at the primary site quiesces applications running thereat for performing error recovery procedures and signalling both primary and secondary locations that failed duplex has occurred. The error recovery program determines the cause of failed duplex, and if error recovery is successful, duplex mode is resumed. If the error recovery program is unable to perform error recovery successfully, then further write updates at the primary site are prohibited and an error message is communicated to operators at both the primary and secondary sites.
摘要:
A method and system are disclosed for permitting high concurrency of access during backup copying of designated data stored within a storage subsystem which includes multiple storage devices coupled to the data processing system via a storage subsystem control unit having subsystem memory therein. Data within each storage device is accessed utilizing a Unique Control Block (UCB), which identifies a selected storage device, and an associated data retrieval command sequence which identifies the data to be accessed. Portions of the data copied to subsystem memory within the subsystem storage control unit as sidefiles from a first storage device may be accessed utilizing a Unique Control Block (UCB) associated with an alternate storage device by associating a selected data retrieval command sequence therewith which identifies the data as stored within the subsystem memory. In this manner, data from a first storage device may be accessed utilizing an alternate Unique Control Block (UCB), permitting concurrent access of data utilizing the Unique Control Block (UCB) associated with the first storage device.
摘要:
A method and system are disclosed for enhanced efficiency of backup copying of designated datasets stored within a plurality of storage devices coupled to the data processing system via a storage subsystem control unit having subsystem memory therein. Application execution within the data processing system is temporarily suspended long enough to form a dataset logical-to-physical system address concordance to be utilized to administer copying of the designated dataset. Thereafter, application initiated updates to uncopied portions of the designated datasets are temporarily deferred until sidefiles of the affected portions of the designated datasets are written to subsystem memory. The updates are then written to the storage subsystem. Portions of the designated datasets are then accessed and copied from the storage subsystem on a scheduled or opportunistic basis utilizing selected data retrieval command sequences. A sidefile status query is appended to selected data retrieval command sequences and the presence of data within the subsystem memory is determined without the necessity of additional communications between the data processing system and the storage subsystem. The sidefiles are then accessed and copied. Copied portions of the designated datasets and sidefiles are then copied to alternate storage locations in a backup copy order defined by the address concordance.
摘要:
A method in a data processing system for automatically terminating or resuming backup copy sessions after an abnormal interrupt or reset notification occurrence during a backup copy process. A status indication is entered into activity tables associated with a plurality of storage subsystems and devices within a data processing system in response to initiation of a backup copy session. Status indications are then entered upon successful completion of a backup copy session within the data processing system. The indications within the activity tables are reviewed to determine the status of a backup copy session upon restarting a resource manager, abnormal termination of a backup copy program, or an operating system initial program load. If a backup copy session has been initiated but not completed, the backup copy session is then terminated. The indications within the activity tables are also reviewed to determine the status of a backup copy session if a reset notification is raised by a storage subsystem. The tracks extents which are active for a volume associated with a particular session identification are determined. A comparison is then made between the tracks extents which are active and the volume and extent information associated with a physical session identification. If a match exists between the tracks extents which are active and the volume and extent information associated with a physical session identification, the backup copy session resumes. If a match does not exist, the backup copy session is terminated.
摘要:
A method, apparatus and program storage device for providing asynchronous status messaging in a data storage system Asynchronous events are returned from a target to an initiator in response to a Read Message from the initiator. The return status message may or may not be associated with a previously given command or Logical Unit (LUN).
摘要:
A method to initialize information disposed in an information storage and retrieval system comprising a data cache and one or more information storage media comprising a plurality of addresses. The method receives an initialization command, where that initialization command comprises a range of addresses and, optionally, an initialization pattern, where the plurality of addresses includes the range of addresses. The method forms and saves state information, where that state information includes the initialization pattern, and generates and saves a plurality of indicators, where that plurality of indicators includes an indicator for each track comprising the range of addresses. The method sets each of those indicators to a first value thereby indicating special handling, and provides an ending status signal, such that the receiving step, forming step, generating step, setting step, and providing step are performed substantially synchronously.
摘要:
A method and apparatus for improving storage controller throughput when servicing non-specific read requests. Throughput is increased by enabling the storage controller to service non-specific read requests issued against a device address which is not the same as the device address associated with the data that the non-specific read is requesting. The device address associated with the data is thus free to be used for another I/O operation. When the device address used by the non-specific read is a non-busy type (that is, the device and/or data is not locked) then the throughput improvement is obtained by increasing the number of data requests that may be concurrently handled by the storage controller.
摘要:
A method and system are disclosed for transferring backup copies of designated multi-track datasets which are stored in a plurality of storage subsystems to a host system via data channel during concurrent execution of an application. Upon receiving a request for a backup copy, execution of an application is temporarily suspended. A dataset logical-to-physical storage system address concordance is formed and execution of the application is resumed thereafter. Any application initiated updates to affected tracks within the designated datasets which occur prior to generation of a backup copy are processed by buffering the updates, copying the affected tracks to storage subsystem memory, updating the address concordance and writing the updates to the affected tracks within the dataset. Requests from the host system to transfer a selected sequence of tracks within the designated multi-track dataset are then examined to determine if the sequence includes affected tracks which have previously been copied to storage subsystem memory. An artificial track is then automatically substituted into the requested sequence of selected tracks for each affected track which has previously been copied to storage subsystem memory during transfer of the selected sequence of tracks such that the host system is made aware of the previous copying of these affected tracks.