[ 
https://issues.apache.org/jira/browse/CASSANDRA-13750?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16119713#comment-16119713
 ] 

Aleksey Yeschenko commented on CASSANDRA-13750:
-----------------------------------------------

If you have legacy pre-2.1 data shards lying around in sstables, this bug will 
hurt. One minor problem is that for people who don't, there will be a short 
period of digest mismatches during the minor upgrade from 3.0.prev to 3.0.next, 
but I don't see a way to avoid it.

+1

> Counter digests include local data
> ----------------------------------
>
>                 Key: CASSANDRA-13750
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-13750
>             Project: Cassandra
>          Issue Type: Bug
>            Reporter: Blake Eggleston
>            Assignee: Blake Eggleston
>            Priority: Minor
>             Fix For: 4.0, 3.0.x, 3.11.x
>
>
> In 3.x+, the raw counter value bytes are used when hashing counters for reads 
> and repair, including local shard data, which is removed when streamed. This 
> leads to constant digest mismatches and repair overstreaming.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

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

Reply via email to