seanghaeli commented on code in PR #74522:
URL: https://github.com/apache/airflow/pull/74522#discussion_r4264448752
##########
providers/amazon/src/airflow/providers/amazon/aws/waiters/dms.json:
##########
@@ -141,13 +141,13 @@
"matcher": "path",
"argument": "ReplicationTasks[0].Status",
"expected": "running",
Review Comment:
I've read some DMS docs, and my understanding is that a DMS task has 3
modes: full-load, cdc, and a mix of both. In the later two modes, `running` is
[actually a terminal
state](https://docs.aws.amazon.com/dms/latest/userguide/CHAP_Task.CDC.html#CHAP_Task.CDC.StopPoint)
(unless an extra parameter `CdcStopPosition` is explicitly set), so
instructing retries in this case would just cause it to fail laid when the
retries time out in 30 minutes.
I think what makes the most sense is to set the `state` for both `running`
and `stopping` as `success`. The reason being that `DmsModifyTaskOperator`
requires the task to be `stopped`, `ready`, or `failed` before modifying. The
waiter would only ever see `running` and `stopping` if something started the
task after the modification is finished. If we mark these states as `success`,
the operator will not wait for these states, and will correctly only wait on
`modifying`.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
