摘要:
A map-reduce compatible distributed file system that consists of successive component layers that each provide the basis on which the next layer is built provides transactional read-write-update semantics with file chunk replication and huge file-create rates. Containers provide the fundamental basis for data replication, relocation, and transactional updates. A container location database allows containers to be found among all file servers, as well as defining precedence among replicas of containers to organize transactional updates of container contents. Volumes facilitate control of data placement, creation of snapshots and mirrors, and retention of a variety of control and policy information. Also addressed is the use of distributed transactions in a map-reduce system; the use of local and distributed snapshots; replication, including techniques for reconciling the divergence of replicated data after a crash; and mirroring.
摘要:
Aspects of the subject matter described herein relate database virtualization. In aspects, clusters of database servers may be located at various locations of the Internet. When a client seeks to access a logical database, the client may send a logical server and logical database name in a data structure. These names may be used to find a physical server(s) and database(s) that correspond to the logical database. Once the location is determined, a proxy component is used to intercept and/or forward communications between the client and the physical server(s) and database(s) corresponding to the logical database. Using this system, a client may access data from a logical database without knowing the physical address of the logical user database.
摘要:
An data processing method includes: determining, from execution-target data, a plurality of data items included in the execution-target data, a same series of a plurality of processes being executed for the plurality of data items; associating identification information with each of the plurality of data items; and executing batch processing of the plurality of data items with which the identification information has been associated.
摘要:
A predicate-based row level security system is used when workers build or split an analytical data store. According to one implementation, predicate-based means that security requirements of source transactional systems can be used as predicates to a rule base that generates one or more security tokens, which are associated with each row as attributes of a dimension. Similarly, when an analytic data store is to be split, build job, user and session attributes can be used to generate complementary security tokens that are compared to security tokens of selected rows. Efficient indexing of a security tokens dimension makes it efficient to qualify row retrieval based on security criteria.
摘要:
A system synchronizes change-data in a multi-tenant system with one or more external service provider systems. The system comprises a system interface configured to receive transaction events; a write-optimized transactional database for recording the transaction events; a replicated read-optimized transactional database; a transaction log of change-data generated from the read-optimized transactional database; a first change-data processor configured to identify the change-data in the transaction log that is relevant to at least one external service provider system, and to generate an aggregate relevant change-data log; a second change-data processor configured to identify aggregate relevant change-data that is relevant to a particular external service provider system, and to generate a service provider relevant change-data log; a service provider connector configure to await a real-time trigger condition, and to, in response to the trigger condition, perform a synchronization session to synchronize the service provider relevant change-data with the particular external service provider system.
摘要:
Techniques are disclosed to transmit arbitrarily large data units for transaction processing in a networked environment. A request is received to store a data unit of a size exceeding an allocated memory address space of a transaction gateway component of the networked environment. A predefined store function, provided by a repository interface component, is invoked to store the data unit to a data repository component of the networked environment and without segmenting the data unit. A repository handle of the stored data unit is identified. A predefined load function, provided by the repository interface component, is invoked to load a portion of the stored data unit, based on the identified repository handle, where the portion is smaller than the stored data unit.
摘要:
A system, method, and computer-readable storage medium having a computer-readable instruction thereon to replicate transactions in a Zettabyte File System (ZFS). The method includes operations including determining, using a tracing module, a commitment of a current file transaction to a disk, obtaining when a probe function associated with the current file transaction is triggered, the current file transaction, recording a data change contained in the current file transaction, registering a callback function and a reference to the data change contained in the current file transaction after the commitment of the current file transaction, and sending the data change to a remote server.
摘要:
A method is used in managing delete operations in files of file systems. An indirect data block of a file of a file system is deleted. A parent indirect data block pointing to the indirect data block is removed from a list. The parent indirect data block has been added to the list for flushing contents of the parent indirect data block to a storage device.
摘要:
Example implementations described herein are directed to a transaction query engine involving a token that is generated by the server and sent to the corresponding the client. The client uses the token to query transactions made on the server. The token can include mountpoint information detailing shared and private information of the filesystems, transaction identifiers identifying transactions of a filesystem, and a filesystem identifier to identify the corresponding filesystem. The server processes this token and updates the token based on the transaction information of the filesystem, and sends the token back to the corresponding client.
摘要:
A technique includes receiving a request to restore a state of a file system. The request defines a rollback period, which extends from a restoration time that is specified by the request to a time at which the request was made. The technique includes reconstructing a state of the file system in response to a query using a preserved snapshot of the file system. The reconstruction including reconstructing the state at an arbitrary query time based at least in part on a dynamically inheritable attribute of an entity stored on the file system and the rollback period.