The Sync mechanism synchronizes data between data sources and the Fabric database. During a synchronization, extraction and transformation processes are executed on an LU Instance (LUI) whereby data is retrieved from source systems using the GET command and then loaded into the Fabric database.
Since Fabric receives multiple requests concurrently, the GET command's response time is highly important. To optimize the LUI retrieval process, Fabric uses a cache mechanism that has a predefined size limit which enables loading an instance into the memory faster.
The size and location of the cache on the Fabric server are defined in the [fabricdb] section of the config.ini file.
The size of the cache is limited and can be updated if needed. Note that more disk space is required when the cache size is increased.
The following variables in the [fabricdb] section of the config.ini are used to define the permanent storage and location of the cache in the Fabric server:
MDB_DEFAULT_SCHEMA_CACHE_STORAGE_TYPE, holds the storage type:
#MDB_DEFAULT_SCHEMA_CACHE_STORAGE_TYPE=CASSANDRA
When the storage type in the config.ini is set to NONE, the instance is not saved to Fabric storage.
As experimental feature on Fabric 6.5.2, it is also possible to define S3 or AZURE_BLOB_STORE as the storage type.
MDB_DEFAULT_CACHE_PATH, holds the location of the Fabric MDB cache files on the server.
#MDB_DEFAULT_CACHE_PATH=/dev/shm/fdb_cache
MDB_DEFAULT_SCHEMA_CACHE_SIZE, defines the size limit of the cache in bytes for MicroDB instances.
Fabric enables you to modify the default location of the cache per Logical Unit using the MDB_DEFAULT_CACHE_PATH variable in the [fabricdb] section of the config.ini and the LU Schema's Cache Location property as follows:
Note that both the config.ini and the Fabric Studio configuration changes become effective only after Fabric is restarted. On the first LU deployment the restart is not required.
The Sync mechanism synchronizes data between data sources and the Fabric database. During a synchronization, extraction and transformation processes are executed on an LU Instance (LUI) whereby data is retrieved from source systems using the GET command and then loaded into the Fabric database.
Since Fabric receives multiple requests concurrently, the GET command's response time is highly important. To optimize the LUI retrieval process, Fabric uses a cache mechanism that has a predefined size limit which enables loading an instance into the memory faster.
The size and location of the cache on the Fabric server are defined in the [fabricdb] section of the config.ini file.
The size of the cache is limited and can be updated if needed. Note that more disk space is required when the cache size is increased.
The following variables in the [fabricdb] section of the config.ini are used to define the permanent storage and location of the cache in the Fabric server:
MDB_DEFAULT_SCHEMA_CACHE_STORAGE_TYPE, holds the storage type:
#MDB_DEFAULT_SCHEMA_CACHE_STORAGE_TYPE=CASSANDRA
When the storage type in the config.ini is set to NONE, the instance is not saved to Fabric storage.
As experimental feature on Fabric 6.5.2, it is also possible to define S3 or AZURE_BLOB_STORE as the storage type.
MDB_DEFAULT_CACHE_PATH, holds the location of the Fabric MDB cache files on the server.
#MDB_DEFAULT_CACHE_PATH=/dev/shm/fdb_cache
MDB_DEFAULT_SCHEMA_CACHE_SIZE, defines the size limit of the cache in bytes for MicroDB instances.
Fabric enables you to modify the default location of the cache per Logical Unit using the MDB_DEFAULT_CACHE_PATH variable in the [fabricdb] section of the config.ini and the LU Schema's Cache Location property as follows:
Note that both the config.ini and the Fabric Studio configuration changes become effective only after Fabric is restarted. On the first LU deployment the restart is not required.