You are viewing a plain text version of this content. The canonical link for it is here.

Posted to commits@cassandra.apache.org by "Jonathan Ellis (JIRA)" <ji...@apache.org> on 2013/10/13 20:31:45 UTC

[jira] [Resolved] (CASSANDRA-6192) Ignore gc_grace when all replicas have ACKed the delete

     [ https://issues.apache.org/jira/browse/CASSANDRA-6192?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Jonathan Ellis resolved CASSANDRA-6192.
---------------------------------------

    Resolution: Duplicate

See comments on 3620 starting with https://issues.apache.org/jira/browse/CASSANDRA-3620?focusedCommentId=13561779&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-13561779

> Ignore gc_grace when all replicas have ACKed the delete
> -------------------------------------------------------
>
>                 Key: CASSANDRA-6192
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-6192
>             Project: Cassandra
>          Issue Type: Wish
>          Components: Core
>            Reporter: André Cruz
>            Priority: Critical
>
> When a client issues a delete with a consistency level >= QUORUM, all replicas are contacted, even though the coordinator may return the answer before all responses arrive if QUORUM. Therefore, in the usual case when all replicas are alive, the coordinator will know when a delete has been ACKed by all replicas responsible for that data. In this situation I think it would be beneficial if the coordinator could notify the replicas that that tombstone is safe to purge on the next compaction, regardless of the gc_grace value.
> This would make tombstones disappear much faster than they normally would.



--
This message was sent by Atlassian JIRA
(v6.1#6144)