You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@cassandra.apache.org by "Randall Leeds (JIRA)" <ji...@apache.org> on 2010/11/12 23:52:14 UTC

[jira] Updated: (CASSANDRA-982) read repair on quorum consistencylevel

     [ https://issues.apache.org/jira/browse/CASSANDRA-982?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Randall Leeds updated CASSANDRA-982:
------------------------------------

    Attachment: 0001-better-digest-checking-for-ReadResponseResolver.patch

Agree that more could be done here to aggressive resolve difference with the responses that come in after the quorum. For this reason I'd recommend leaving this ticket open, but I think this patch is applicable here and not worth a separate ticket.

Right now the digest checking is such that the ReadResponseResolver might miss an inconsistency.
Rather than comparing the digests of all the digest requests to one another, the last one seen "wins" and is compared to the digest of each version seen from a data request.

This patch catches this case and additionally fast-tracks the exception by catching it as we iterate through the responses rather than re-iterating the versions afterward.

> read repair on quorum consistencylevel
> --------------------------------------
>
>                 Key: CASSANDRA-982
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-982
>             Project: Cassandra
>          Issue Type: New Feature
>          Components: Core
>            Reporter: Jonathan Ellis
>            Assignee: Matthew F. Dennis
>            Priority: Minor
>             Fix For: 0.7.1
>
>         Attachments: 0001-better-digest-checking-for-ReadResponseResolver.patch
>
>
> CASSANDRA-930 made read repair "fuzzy optional," but this only helps with ConsistencyLevel.ONE:
> - Quorum reads always send requests to all nodes
> - only the first Quorum's worth of responses get compared
> So what we'd like to do two changes:
> - only send read requests to the closest R live nodes
> - if read repair is enabled, also compare results from the other nodes in the background

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.