[jira] [Commented] (CASSANDRA-9038) Atomic batches and single row atomicity appear to have no test coverage

["Edward Capriolo (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CASSANDRA-9038?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15526968#comment-15526968 ] 

Edward Capriolo commented on CASSANDRA-9038:
--------------------------------------------

[~aweisberg] I have just done some work to demonstrate using black-box testing some aspects of how batches work. 

https://github.com/edwardcapriolo/ec/tree/master/src/test/java/Base/batch

The challenge as I see it, how does one prove atomicity. Trying an operation N thousand times and asserting a result is slow and not really a "proof" of anything. 

(I am not a fan of mockito) but I would say the only thing we can assert or prove: 
* If we were able to place mocks in Memtables
* issue statements like batch-mutations, possibly a batch across 10 row keys
* assert that specific methods were called a specific number of times

Being that I am not very familiar with any of the api's this could be a vast oversimplification of the process. Maybe someone else can chime in and say what specifically we could do.My worry is to add some mock test that always passes because of a limited scope and is only a burden that has to be changed with api changes.

> Atomic batches and single row atomicity appear to have no test coverage
> -----------------------------------------------------------------------
>
>                 Key: CASSANDRA-9038
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-9038
>             Project: Cassandra
>          Issue Type: Test
>            Reporter: Ariel Weisberg
>
> Leaving the solution to this up to the assignee. It seems like this is a guarantee that should be checked.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)