Github user zentol commented on a diff in the pull request: https://github.com/apache/flink/pull/6112#discussion_r192666792 --- Diff: docs/ops/upgrading.md --- @@ -172,7 +172,7 @@ First major step in job migration is taking a savepoint of your job running in t You can do this with the command: {% highlight shell %} -$ bin/flink savepoint :jobId [:targetDirectory] +$ ./bin/flink savepoint <jobId> [savepointDirectory] --- End diff -- `savepointDirectory` has specific semantics in the documentation and is assumed to be the directory with which one can start a job from again. The directory passed here _does not fit these semantics_ as a `savepointDirectory` is created *within* the `targetDirectory`. Please revert.
---