"End task" very often does not end the task. I'm not sure what it actually does, but much of the time it just ignores the request for some long period of time.
- What is the reason for this?
- Can a way be added to immediately kill the task/process?

What kind of Task type are we talking about here?

---

We notice this with SQL tasks frequently. But I'm sure its happened with other types as well.

It is different for each Task type as different API:s allow different ways of aborting. It is also related to where in the execution process the Task is. So, sometimes it is in a state where it cannot be aborted but need to finish a certain part first.

---

Can't you always just execute the task in a separate process and kill that process when End Task is invoked? Whatever the task was running needs to be able to handle this kind of aborting anyway.

It seems to me like End Task is mixing two different pieces of functionality:
- sending an abort request- which can be handled by applications gracefully that support such a request
- "pulling the plug" on an application (immediately killing the process- similar to what you might do with "End Task" in Task Manager in Windows)

We rarely ever want the first one of these. We almost always want the second.

In my specific case about SQL tasks- SQL Server will perform a rollback regardless of whether the calling task is still live- so I don't see a reason not to immediately kill the entire calling process.

Can't you always just execute the task in a separate process and kill that process when End Task is invoked? Whatever the task was running needs to be able to handle this kind of aborting anyway.

It seems to me like End Task is mixing two different pieces of functionality:
- sending an abort request- which can be handled by applications gracefully that support such a request
- "pulling the plug" on an application (immediately killing the process- similar to what you might do with "End Task" in Task Manager in Windows)

We rarely ever want the first one of these. We almost always want the second.

In my specific case about SQL tasks- SQL Server will perform a rollback regardless of whether the calling task is still live- so I don't see a reason not to immediately kill the entire calling process.

No, because that would create to much overhead. Both in terms of launching own exe and the communication between the exe and the main service.

---