At present, the reload operation is executed serially: if there are multiple configurations change, the effective time of the change will be relatively long due to the serial operation of the interface; at the same time, the reload operation requires external batch loading according to the current cluster reload status, and these settings are aggravated considering the complexity of interface usage. so I consider making the following adjustments:
1. The reload operation is changed to task submission mode. After receiving the task, the Master returns a task ID to the caller; the caller queries the task completion status through the task ID;
2. The Master controls the batches loaded and the number of each load; and periodically records the execution results of reload in the task ID.
At present, the reload operation is executed serially: if there are multiple configurations change, the effective time of the change will be relatively long due to the serial operation of the interface; at the same time, the reload operation requires external batch loading according to the current cluster reload status, and these settings are aggravated considering the complexity of interface usage. so I consider making the following adjustments:
1. The reload operation is changed to task submission mode. After receiving the task, the Master returns a task ID to the caller; the caller queries the task completion status through the task ID;
2. The Master controls the batches loaded and the number of each load; and periodically records the execution results of reload in the task ID.
JIRA link - [INLONG-386] created by gosonzhang