摘要:
An apparatus, system, and method are disclosed for a read-before-write storage controller instruction. A sequencer receives an atomic read-before-write instruction comprising new data, a target address for the new data, and an undo log address. An I/O unit reads old data from the target address, writes the old data and the target address to the undo log address, and writes the new data to the target address as directed by the sequencer.
摘要翻译:公开了一种读写前存储控制器指令的装置,系统和方法。 定序器接收包括新数据,新数据的目标地址和撤销日志地址的原子预写指令。 I / O单元从目标地址读取旧数据,将旧数据和目标地址写入撤销日志地址,并按照定序器的指示将新数据写入目标地址。
摘要:
An apparatus, system, and method are disclosed for a read-before-write storage controller instruction. A sequencer receives an atomic read-before-write instruction comprising new data, a target address for the new data, and an undo log address. An I/O unit reads old data from the target address, writes the old data and the target address to the undo log address, and writes the new data to the target address as directed by the sequencer.
摘要翻译:公开了一种读写前存储控制器指令的装置,系统和方法。 定序器接收包括新数据,新数据的目标地址和撤销日志地址的原子预写指令。 I / O单元从目标地址读取旧数据,将旧数据和目标地址写入撤销日志地址,并按照定序器的指示将新数据写入目标地址。
摘要:
Provided are a method, system, and program for managing a relationship between one target volume and one source volume. For each of the source volume and target volume, the memory includes: (i) at least one element, wherein each element represents a range of sequential data units in the volume; (ii) at least one relationship entry, wherein each relationship entry represents one relationship; (iii) at least one element pointer associating one element with one relationship entry, wherein the data units represented by the element are part of the relationship represented by the relationship entry that the pointer associates with the element; and (iv) one relationship pointer for each relationship entry associating the relationship entry with volume metadata, wherein the volume metadata provides information on the relationship represented by the relationship entry.
摘要:
Provided are a method, system, and article of manufacture wherein a command is received for writing data to a first storage location. A determination is made that previously written data is stored in the first storage location. The previously written data is copied to a second storage location, in response to determining that the previously written data is stored in the first storage location. The data is written to the first storage location, in response to copying the previously written data to the second storage location.
摘要:
Provided are a method, system, and program for managing a relationship between one target volume and one source volume. Information is maintained in memory on an existing relationship between at least one source volume and at least one target volume, comprising: (i) at least one element, wherein each element represents a range of sequential data units in the volume; (ii) at least one relationship entry, wherein each relationship entry represents one relationship; and (iii) at least one element pointer associating one element with one relationship entry, wherein the data units represented by the element are part of the relationship represented by the relationship entry that the pointer associates with the element. A new relationship between at least one target volume and at least one source volume is added. Further, added is a new relationship entry in the memory representing the new relationship and one new element pointer is added in the memory for each element including data units included in the new relationship to associate the element with the new relationship entry, wherein the data units represented by the element are part of the new relationship.
摘要:
A method that can simplify a recovery of a system. The method includes storing multiple types of information, and includes: a first stage of storing information representative of a content of a persistent memory entity at a certain point in time; and a second stage of storing information representative of a state of a virtual machine at the certain point in time; and recovering a system that associates between the first stage of storing information and the second stage of storing information.
摘要:
An apparatus for optimizing a transaction comprising an initial sequence of computer operations, the apparatus includes a processing unit which identifies one or more idempotent operations comprised within the initial sequence, and which reorders the initial sequence to form a reordered sequence comprising a first sub-sequence of the computer operations followed by a second sub-sequence of the computer operations, the second sub-sequence comprising only the one or more idempotent operations.
摘要:
A method for managing a data storage system includes copying data stored on a first storage subsystem to a second storage subsystem in an asynchronous mirroring process. A record is maintained on the second storage subsystem, indicative of locations at which the data have been updated on the first storage subsystem and have not yet been copied to the second storage subsystem. Upon receiving at the second storage subsystem, from a host processor, a request to access the data stored at a specified location on the data storage system, if the specified location is included in the record, the second storage subsystem initiates a synchronous transfer of the data at the specified location from the first storage subsystem.
摘要:
A method for storing multiple types of information, the method includes: storing information representative of a content of a persistent memory entity at a certain point in time; and storing information representative of a state of a virtual machine at the certain point in time. A computer program product including a computer usable medium including a computer readable program, wherein the computer readable program when executed on a computer causes the computer to: store information representative of a content of a persistent memory entity at a certain point in time; and store information representative of a state of a virtual machine at the certain point in time.
摘要:
Disclosed is a technique for applying changes back to an original copy of data to reduce an amount of data transferred. A first indicator is maintained for each block of data in a first storage to indicate whether the block of data has been updated in the first storage since the block of data was last transferred to a second storage. A second indicator is maintained for each block of data in the second storage to indicate whether the block of data has been updated in the second storage since the block of data in the second storage was overwritten by a corresponding block of data in the first storage. Upon receiving a reverse restore operation, each block of data in the second storage for which a second indicator has been set to indicate that the block of data has been updated is transferred to the first storage and each block of data in the second storage that corresponds to a block of data in the first storage for which a first indicator has been set to indicate that the block of data in the first storage has been updated is transferred to the first storage.