sjwiesman commented on issue #8861: [FLINK-12963][state-processor-api] Add 
savepoint writer for bootstrapping new savepoints
URL: https://github.com/apache/flink/pull/8861#issuecomment-505953781
 
 
   Thanks for taking a look @Myasuka 
   
   Parallelism does not matter since it does not affect the user's ability to 
restore, only max parallelism which is used to define key groups. This 
implementation assumes the parallelism of the provided data set was set 
appropriately by the user for the data. The only times we change it are when 
   
   1) Writing broadcast state
   2) The parallelism is greater than the specified max parallelism
   
   As outlined in the docs, max parallelism is set either when creating a new 
savepoint `Savepoint.create` or when using an existing savepoint as a base 
`Savepoint.load` its max parallelism is used.
   
   I do agree, the ability to set max functionality at a per operator level, in 
addition to the global max parallelism would be useful. 

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