[GitHub] [cassandra] yifan-c commented on pull request #1168: CASSANDRA-16878 trunk: Race in commit log replay can cause rejected mutations

[GitBox <gi...@apache.org>]

yifan-c commented on pull request #1168:
URL: https://github.com/apache/cassandra/pull/1168#issuecomment-1018086994


   Hi, @adelapena. Firstly, thanks for the patch. 
   
   If my understanding is correct, the fix is to divide the mutations by schema change into multiple groups. What concerns me is that the schema change is replayed along with the other mutations in the group in parallel. It is possible that some mutations gets dropped because the schema change is replayed out of order or take more time to be applied. 
   
   I have a hacky reproduction that test replay a create table mutation followed with mutations to insert rows. The subsequent mutations are expected to be applied, but they are in fact not. See https://github.com/yifan-c/cassandra/commit/81b52b42e3b6db99989664323790b782aafdf9c5
   
   I think the schema change should be in its own group. 


-- 
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.

To unsubscribe, e-mail: pr-unsubscribe@cassandra.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org



---------------------------------------------------------------------
To unsubscribe, e-mail: pr-unsubscribe@cassandra.apache.org
For additional commands, e-mail: pr-help@cassandra.apache.org