zhijiangW edited a comment on issue #9132: [FLINK-13245][network] Fix the bug 
of file resource leak while canceling partition request
URL: https://github.com/apache/flink/pull/9132#issuecomment-515464324
 
 
   After some offline discussions with @StephanEwen  and @azagrebin . We could 
make the current netty stack logic simple for always calling 
`notifySubpartitionConsumed` while releasing partition's reader/view in any 
cases. Considering the future possible improvements related to network stack, 
we would further confirm what should be done then after deciding to do and 
confirming the goal. 
   
   For current pipelined mode, the current assumption is that one subpartition 
would only have one consumption attempt. 
   
   For current blocking mode, one subpartition could have multiple 
readers/views and the partition release is up to scheduler/JM.
   
   In release-1.10 we could further refactor the current process as follows:
   
   - In netty handler stack, it would only call view/reader.releaseResources in 
the current cases.
   
   - In reader/view stack, it would further notify the parent of the released 
state, maybe still via the method of `notifySubpartitionConsumed`.
   
   - In partition stack, it could decide whether to release itself or not based 
on some factors. E.g. for pipelined partition it could release partition as 
well. And for blocking partition it would do nothing.
   

----------------------------------------------------------------
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.
 
For queries about this service, please contact Infrastructure at:
[email protected]


With regards,
Apache Git Services

Reply via email to