Ngone51 commented on issue #25577: [WIP][CORE][SPARK-28867] InMemoryStore 
checkpoint to speed up replay log file in HistoryServer
URL: https://github.com/apache/spark/pull/25577#issuecomment-533847127
 
 
   >  It's not even verified via POC, so I can't see any reason to adopt it 
right now.
   
   Oh, sorry. I was just discussing the possible integrate way in the following 
design and not ask you to  make the change right now. 
   
   > The thing is, caller logic should be different based on the support of 
sync/async snapshot. We'll have dump()/dumpAsync() and supportAsyncDump() and 
let caller deal with this. 
   
   Sorry, I don't know this part well. Why do we need to implement sync and 
async dump separately ? And what's factor should caller depends on to decide 
which way it should choose ? If we have both sync and async way to dump 
KVStore, why don't we just choose async way to reduce latency ?
   
   I always think that KVStore should just implement dump() and hide the 
details from the caller, e.g. whether it's sync or async. And KVStore should 
always expose  its interface to caller as a single view without considering it 
supports sync dump or async dump while using it.
   
   
    

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

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to