Data

Dragon has APIs for managing data in a scalable fashion. A prominent member of these APIs is the distributed dictionary. Descriptions and APIs are grouped together for each of the supported Dragon Data Types.

Python Reference

In-Memory Distributed Dictionary

DDict

The Distributed Dictionary provides a key/value store that is distributed across a series of managers and on nodes of a Dragon run-time.

DDictError

This is the generic error that all other Distributed Dictionary specific

DDictSyncError

DDictFullError

DDictTimeoutError

DDictPersistCheckpointError

DDictKeyError

DDictCheckpointSyncError

DDictUnableToCreateError

DDictFutureCheckpointError

DDictManagerStats

Included in manager stats are the manager identifier (0 to num_managers-1), the total number of bytes in the manager's pool, the total used bytes in the manager's pool, the number of key/value pairs stored in the manager, and the dictionary of free blocks.

CheckpointPersister

This class declares methods that all checkpoint persistence classes should implement.

NULLCheckpointPersister

PosixCheckpointPersister

DAOSCheckpointPersister

DDictKeysView

A live view of the keys of a DDict object.

DDictValuesView

A live view of the values of a DDict object.

DDictItemsView

A live view of the items of a DDict object.

FilterContextManager

A context manager that cleans up the filtering processes upon exit.

Zarr Store

Store

This class implements a Zarr store using the Dragon distributed dictionary (DDict).

C Reference

C++ Reference