In case of need, users can adjust the environment variables in this table to improve performance.
For most use cases, the default setting is optimal.
|MAX_NODE_ID||This variable controls the random allocation of IDs to vworkers.
It must not be smaller than the total number of vworkers in the cluster, which is the number of nodes multiplied by the number of vworkers per node. And it has a range from 50 to 999.
The default setting is large enough for standard Aster Database configurations of 16 worker nodes.
Adjusting the value for this variable may improve performance for small clusters when establishing a connection to the database using the AsterR package, and is required if the cluster has more than 16 worker nodes.
|SERVER_STREAM_DIR||This variable sets the directory on the Aster cluster to store virtual object files.||/home/extensibility/stream|
|STREAM_READ_BATCH_CNT||This variable controls the number of records to read as a batch for stream processing.
The default is all records read at once.
Adjusting the value for this variable if the input data stream is larger than the worker node memory.
When this variable is set to a non-default value, the operation is performed independently for each batch of stream data. This is not recommended for aggregate operations such as training a model.
|STREAM_MEMORY_LIMIT_MB||This variable controls the amount of memory allocated to the stream for the execution of a R function using aa.apply and aa.tapply.
If the system reports memory allocation errors, try increasing the value of this variable.
Commands to display and change the environment variables
To display the value of an environment variable:sys.getenv(“STREAM_MEMORY_LIMIT_MB”)
To change the value of an environment variable:
sys.getenv(STREAM_MEMORY_LIMIT_MB = 20000)