zentol opened a new pull request #9296: [FLINK-13476][coordination] Release 
pipelined partitions on vertex restart / job failure
URL: https://github.com/apache/flink/pull/9296
 
 
   Based on #9250.
   
   Only commits tagged for FLINK-13476 are relevant.
   
   Re-introduces a partition release call for pipelined partitions in case 
`Execution#cancel` is called after a vertex has been FINISHED. Since a task 
producing a pipelined partition may finish before any consumer ever attempted 
to consume the partition, it can happen that said partition is never release 
until the TE shuts down.
   That is because the partition lifecycle management only cares about blocking 
partitions.
   
   Note that cancel() is not only called for actually canceling the execution, 
but _always_ when a job fails or if the containing failover region is being 
restarted. This covers all cases where a pipelined partition can be produced, 
but may never be consumed.

----------------------------------------------------------------
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