Use SHOW DDL to view DDL jobs - PolarDB

In PolarDB-X, every DDL statement runs as a DDL job. Use SHOW DDL to inspect active and recent DDL jobs in your instance.

Syntax

SHOW [FULL] DDL;

Example

The following example shows the output of SHOW DDL\G for a paused ALTER TABLE job. The BACKFILL_PROGRESS and PHY_DDL_PROGRESS fields reflect different internal sub-phases, so their values can differ from the overall PROGRESS.

SHOW DDL\G;
*************************** 1. row ***************************
           JOB_ID: 1359947811109470208
    OBJECT_SCHEMA: d1
      OBJECT_NAME: t2
           ENGINE: DAG
         DDL_TYPE: ALTER_TABLE
            STATE: PAUSED
BACKFILL_PROGRESS: 0%
 PHY_DDL_PROGRESS: 100%
         PROGRESS: 20%
       START_TIME: 2021-08-05 11:01:25.291
         END_TIME: 2021-08-05 11:02:27.020
 ELAPSED_TIME(MS): 61729
      PHY_PROCESS:
       CANCELABLE: true
1 row in set (0.03 sec)

Response parameters

Parameter	Description
`JOB_ID`	The unique ID of the DDL job. A 64-bit signed integer of the `LONG` data type.
`OBJECT_SCHEMA`	The schema name of the object on which the DDL job is performed.
`OBJECT_NAME`	The name of the object on which the DDL job is performed — for example, the table name for a `CREATE TABLE` or `ALTER TABLE` job.
`ENGINE`	The engine type for the DDL job. Default value: `DAG`.
`DDL_TYPE`	The type of the DDL job, such as `CREATE_TABLE` or `ALTER_TABLE`.
`STATE`	The current state of the DDL job. Valid values: `QUEUED`, `RUNNING`, `COMPLETED`, `ROLLBACK_RUNNING`, `ROLLBACK_COMPLETED`, `PAUSED`. See DDL job states for state transitions.
`BACKFILL_PROGRESS`	The progress of data backfilling within the DDL job.
`PHY_DDL_PROGRESS`	The progress of batch physical DDL statements being executed within the DDL job.
`PROGRESS`	The overall progress of the DDL job.
`START_TIME`	The time at which the DDL job started.
`END_TIME`	The time at which the DDL job ended.
`ELAPSED_TIME(MS)`	The time elapsed since the DDL job started, in milliseconds.
`PHY_PROCESS`	The state of physical DDL statements currently being executed.
`CANCELABLE`	Whether the DDL job can be canceled using `CANCEL DDL`. `true` means the job is cancelable.

DDL job states

The STATE field reflects the current phase of a DDL job. The following diagram shows how states transition:

Under normal conditions, a DDL job progresses through: QUEUED → RUNNING → COMPLETED.

If an exception occurs, the job may be automatically rolled back: QUEUED → RUNNING → ROLLBACK_RUNNING → ROLLBACK_COMPLETED.

What's next

To cancel a cancelable DDL job (CANCELABLE: true), run CANCEL DDL.