Note
This view is available starting with V4.4.1.
Purpose
The DBA_OB_VECTOR_INDEX_TASKS view displays information about all vector index partition rebuild tasks, segment merge tasks, and segment merge dump tasks in the current tenant. You can identify the task type by checking the TASK_TYPE column.
Columns
| Column | Data type | Nullable | Description | |
|---|---|---|---|---|
| TABLE_ID | bigint(20) | YES | The ID of the table. | |
| TABLET_ID | bigint(20) | YES | The ID of the partition. | |
| TASK_ID | bigint(20) | YES | The ID of the task. | |
| START_TIME | timestamp(6) | YES | The start time of the task. | |
| MODIFY_TIME | timestamp(6) | YES | The end time of the task. | |
| TRIGGER_TYPE | varchar(16) | NO | The type of the task trigger. Valid values:
|
|
| STATUS | varchar(16) | NO | The status of the task. Valid values:
|
|
| TASK_TYPE | bigint(20) | NO | The type of the task. Valid values:
|
|
| TASK_SCN | bigint(20) | NO | The version number of the task. | |
| RET_CODE | bigint(20) | NO | The status of the task at the end of the task. Valid values:
|
|
| TRACE_ID | varchar(512) | NO | The trace ID of the task. | |
| PROGRESS_INFO | varchar(512) | NO | The progress information of the task in JSON format. The information includes the current operation, estimated number of rows, completed number of rows, completion percentage, and estimated remaining time. The information is updated every 10 seconds. The JSON string contains the following fields:
|
|
| TASK_INFO | long text | NO | The detailed progress information of the task in JSON format. The information includes the details of each segment for the segment merge task, such as the SCN, key range, index, and memory usage. This information helps you comprehensively understand the execution process of the task. |
Sample query
Query the information about all vector index performance optimization tasks of all tenants in the sys tenant.
SELECT * FROM oceanbase.DBA_OB_VECTOR_INDEX_TASKS ORDER BY MODIFY_TIME DESC\G;
The query result is as follows:
*************************** 1. row ***************************
TABLE_ID: 500178
TABLET_ID: 1152921504606847126
TASK_ID: 14556
CREATE_TIME: 2026-03-25 23:42:43.330766
MODIFY_TIME: 2026-03-25 23:42:43.332887
TRIGGER_TYPE: USER
STATUS: RUNNING
TASK_TYPE: 5
TASK_SCN: -1
RET_CODE: -1
TRACE_ID: Y2C277F000001-00064DDA41D11952-0-1
PROGRESS_INFO: NULL
TASK_INFO: NULL
EXEC_ADDR: NULL
PRIORITY: 0
START_TIME: NULL
END_TIME: NULL
ERR_MSG: NULL
1 row in set
The following table describes the JSON string of the TASK_INFO column.
| Column | JSON field type | Description |
|---|---|---|
| merge_segments_cnt | string | The number of segments to be merged. |
| merge_segments | array | The information of the segments to be merged. The number of elements in the array is the same as that in the merge_segments_cnt column. |
| result_segments | array | The information of the merged segments. |
| scn | string | The internal timestamp. |
| start_key | string | The upper bound of the key range of the segment. |
| end_key | string | The lower bound of the key range of the segment. |
| index_type | string | The type of the index used to process the vectors of the segment. Valid values:
|
| mem_used | string | The memory used by the segment. |
| min_vid | string | The minimum vector ID of the segment. |
| max_vid | string | The maximum vector ID of the segment. |
References
- DBA_OB_VECTOR_INDEX_TASKS
- For more information about the relationship between HNSW vector index dump, segment merge, and task types, see Vector index principles.
- For more information about O&M and manual triggers, see Monitor and maintain vector indexes