In a topology task, you can view the status of each task node to quickly locate execution problems. You can also manually intervene in the task's execution.
Task node statuses
The lifecycle of a task node is as follows: 
A task node in a topology task can be in one of the following states:
Not executed: The node has not been instantiated.
Running: The node is running.
Succeeded: The node ran successfully. Subsequent nodes are automatically triggered unless you manually intervene.
Failed: The node failed to run. All subsequent nodes are not triggered. The entire topology task fails.
Paused: A running node enters this state when you manually pause it. When you resume a paused node, it continues from the breakpoint.
Stopped: The node was forcibly stopped.
Skipped: A failed or stopped node can be marked as skipped to allow subsequent nodes in the topology to run.
Waiting: If a node is set to wait for an event, it enters the waiting state after the preceding node completes. It changes to the running state after the delay time is reached.
Manual intervention operations
You can manually intervene in a running topology task in real time to change its execution logic. For example, you can pause an abnormal node to inspect the issue and then resume execution from the breakpoint. You can also use lock and unlock operations to require manual confirmation for a task node.
Operation type | Object | Description | Applicable task status |
|---|---|---|---|
Pause | Node | Use this operation to gracefully pause a task when you find an abnormality during execution. After you fix the issue, you can resume the task. The corresponding operation is Resume. Important
| Running |
Resume | Node | Continue running a paused task. | Paused |
Stop | Node | Forcibly stop a task that you no longer need to run. |
|
Retry | Node | Rerun the task node from a failed scheduling record. |
|
Skip | Node | Skip the current node and trigger the execution of subsequent nodes. If the node is locked, you must manually unlock it after skipping it to trigger the subsequent nodes. |
|
Reset | Node | Reset the current task node to the Not executed state and delete any subsequent nodes that have been instantiated. |
|
Retrigger | Node | Retrigger the task node to generate a new scheduling record. This operation applies only to simple job nodes or cluster job nodes. |
|
Lock | Node | After a locked task node runs, it automatically aborts subsequent tasks and waits for manual confirmation. After you manually confirm and unlock the node, the topology continues to run subsequent operations. |
|
Unlock | Node | Remove the lock. If the current node is in the Succeeded or Skipped state, subsequent nodes are automatically triggered after you unlock it. | Locked |
Pause All | Topology | Pause all running cluster job nodes and simple job nodes (that contain multiple steps and have not reached the final step) in the current topology task. | Running |
Cancel All | Topology | Forcibly stop all running or paused nodes in the current topology task. |
|
Full restore | Topology | Resume a paused topology task. This resumes paused nodes and triggers the subsequent nodes of successfully executed nodes. | Paused |
Child flows support the pause, cancel, and resume operations. Manual intervention in a child flow also affects the execution of the parent topology task.