[
https://issues.apache.org/jira/browse/CASSANDRA-4636?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13468708#comment-13468708
]
Aleksey Yeschenko commented on CASSANDRA-4636:
----------------------------------------------
Now that ABM is the default BATCH behaviour, all dtests that use BATCH queries
are 'atomic batch dtests'.
Here are things I can (and should) test:
- BATCH and UNLOGGED BATCH rejecting counter mutations
- COUNTER BATCH rejecting regular mutations
- UnavailableException if there aren't enough reachable nodes to store the
serialized batch
- absence of UnavailableException in a single-node cluster
- proper acknowledge_by value in different scenarios
(https://issues.apache.org/jira/browse/CASSANDRA-4544)
What I should but can't test deterministically (and tests MUST be
deterministic):
- batch replay in case of coordinator death (there is no way to kill the node
exactly after the batch has been written, but before any mutations have been
applied). Which is a bummer, but I don't see any ways around it.
> Create atomic batch dtests
> --------------------------
>
> Key: CASSANDRA-4636
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4636
> Project: Cassandra
> Issue Type: Sub-task
> Components: Core
> Affects Versions: 1.2.0 beta 1
> Reporter: Jonathan Ellis
> Assignee: Aleksey Yeschenko
> Fix For: 1.2.0
>
>
> Need to add distributed tests: https://github.com/riptano/cassandra-dtest
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
