JOB Statements #
Job 语句用于管理作业的生命周期。
目前 Flink SQL 支持以下 JOB 语句:
- SHOW JOBS
- STOP JOB
执行 JOB 语句 #
以下示例展示如何在 SQL CLI 中执行 JOB 语句.
Flink SQL> SHOW JOBS;
+----------------------------------+----------+---------+-------------------------+
| job id | job name | status | start time |
+----------------------------------+----------+---------+-------------------------+
| 228d70913eab60dda85c5e7f78b5782c | myjob | RUNNING | 2023-02-11T05:03:51.523 |
+----------------------------------+----------+---------+-------------------------+
Flink SQL> SET 'state.savepoints.dir'='file:/tmp/';
[INFO] Execute statement succeed.
Flink SQL> STOP JOB '228d70913eab60dda85c5e7f78b5782c' WITH SAVEPOINT;
+-----------------------------------------+
| savepoint path |
+-----------------------------------------+
| file:/tmp/savepoint-3addd4-0b224d9311e6 |
+-----------------------------------------+
SHOW JOBS #
SHOW JOBS
展示 Flink 集群上的作业。
Attention SHOW JOBS 语句仅适用于 SQL CLI 或者 SQL Gateway.
STOP JOB #
STOP JOB '<job_id>' [WITH SAVEPOINT] [WITH DRAIN]
停止指定作业。
WITH SAVEPOINT
在作业停止之前执行 Savepoin。 Savepoint 的路径可以通过集群配置的
state.savepoints.dir 指定,
或者通过 SET
语句指定(后者有更高优先级)。
WITH DRAIN 在触发 savepoint 之前将 Watermark 提升至最大。该操作会可能会触发窗口的计算。请您注意该操作可能导致您之后从该创建的 savepoint 恢复的作业结果不正确。
Attention STOP JOB 语句仅适用于 SQL CLI 或者 SQL Gateway.