[beam] 01/01: Issue#24161 Updated docstring for Clusters class

[ni...@apache.org]

This is an automated email from the ASF dual-hosted git repository.

ningk pushed a commit to branch Issue#24161
in repository https://gitbox.apache.org/repos/asf/beam.git

commit 000f77fdd0341473ac03cb74134548483760361c
Author: Ning Kang <ni...@gmail.com>
AuthorDate: Thu Nov 17 10:35:53 2022 -0800

    Issue#24161 Updated docstring for Clusters class
    
    Emphasizes that interactive_beam.Clusters class only manages clusters implicitly created by Beam.
    This class is not needed and should not be used otherwise.
---
 .../apache_beam/runners/interactive/interactive_beam.py     | 13 +++++++++++--
 1 file changed, 11 insertions(+), 2 deletions(-)

diff --git a/sdks/python/apache_beam/runners/interactive/interactive_beam.py b/sdks/python/apache_beam/runners/interactive/interactive_beam.py
index d07113fe5e3..d5253f5009e 100644
--- a/sdks/python/apache_beam/runners/interactive/interactive_beam.py
+++ b/sdks/python/apache_beam/runners/interactive/interactive_beam.py
@@ -342,8 +342,17 @@ class Recordings():
 
 
 class Clusters:
-  """An interface to manage clusters running workers that are connected with
-  the current interactive environment.
+  """An interface to control clusters implicitly created and managed by
+  the current interactive environment. This class is not needed and
+  should not be used otherwise.
+  
+  Do not use it for clusters a user explicitly manages: e.g., if you have
+  a Flink cluster running somewhere and provides the flink master when
+  running a pipeline with the FlinkRunner, the cluster will not be tracked
+  or managed by Beam.
+  To reuse the same cluster for your pipelines, use the same pipeline
+  options: e.g., a pipeline option with the same flink master if you are
+  using FlinkRunner.
 
   This module is experimental. No backwards-compatibility guarantees.