You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by GitBox <gi...@apache.org> on 2019/04/26 02:50:02 UTC

[GitHub] [flink] shixiaogang commented on issue #8269: [FLINK-12285][test] Harden savepoint relevant test cases by explicitly canceling test job

shixiaogang commented on issue #8269: [FLINK-12285][test] Harden savepoint relevant test cases by explicitly canceling test job
URL: https://github.com/apache/flink/pull/8269#issuecomment-486906606
 
 
   > most of these should have no positive effect since the `MiniClusterResource` is shut down at the end of the test.
   
   @zentol The problem here is that MiniCluster cannot correctly release memory when it's shutdown. When there exist tasks running in TaskExecutors, TaskExecutors will throw exceptions and will not release direct memory for network buffers and memory manager. Succeeding tests running in the same JVM process may fail due to insufficient amount of direct memory.
   
   The problem can be reproduced when changing the settings of MiniCluster, including the number of network buffers, the number of task managers, the number of slots in task managers, and slot sharing.  
   

----------------------------------------------------------------
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:
users@infra.apache.org


With regards,
Apache Git Services