Introduction

As shown in Figure 1, the whole clustering and archiving process of AscendHac can be divided into five steps in sequence, which are resource initialization, setting and processing of input data, archive running or merging, setting and processing of output data, and resource destruction.

The current clustering supports data input from different sources (such as memory API calling and file reading). Select the data source based on the actual application scenario when setting and processing input data and output data.

Figure 1 Clustering and archiving

Use AscendHac APIs by following the archiving process to prevent unexpected problems or risks.
When using AscendHac APIs, ensure that the steps for resource initialization and destruction are correctly performed to prevent problems such as incorrect resource destruction and memory leakage.
If a non-APP_ERR_OK value is returned when you use AscendHac APIs, rectify this fault by using the AscendHac resource destruction function to prevent unexpected problems or risks during subsequent operations. (AscendHac APIs create objects in singleton mode. Therefore, you need to destroy the existing object before creating a new one.)

Parent topic: AscendHAC API