[ 
https://issues.apache.org/jira/browse/CASSANDRA-18040?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17634542#comment-17634542
 ] 

Caleb Rackliffe commented on CASSANDRA-18040:
---------------------------------------------

+1...both PRs LGTM

We had talked about two others items:

1.) Having a small number of built-in configurations (guessing {{trie}} / 
{{skiplist}} or something similar) that are present even if not specifically 
defined in the YAML.

2.) Cluster-wide (but still locally overridable) changes in DDL-space.

I know we had intended to make #2 a separate Jira, but will #1 be as well?

Thanks!

> Avoid schema mismatch problems on memtable API misconfiguration
> ---------------------------------------------------------------
>
>                 Key: CASSANDRA-18040
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-18040
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: Cluster/Schema
>            Reporter: Branimir Lambov
>            Assignee: Branimir Lambov
>            Priority: Urgent
>             Fix For: 4.1-rc, 4.x
>
>
> The memtable API requires a definition of the memtable configuration to exist 
> in {{{}cassandra.yaml{}}}. If that is not the case when a schema change is 
> received from another node, this can cause the node to reject the schema 
> change and behave incorrectly.
> To fix this, a node should only fail to construct a memtable when it is 
> validating {{CREATE}} or {{ALTER}} statements, but should fall back to 
> default when it receives invalid configuration selections in the schema.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org

Reply via email to