[
https://issues.apache.org/jira/browse/CASSANDRA-14733?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16628435#comment-16628435
]
Benedict commented on CASSANDRA-14733:
--------------------------------------
Ah, thanks for that insight. Perhaps it's better to wait until we fix
monotonic reads with transient replication, then, as at that point we may be
requesting a separate repaired/unrepaired digest as a matter of course.
At the moment, at least transient requests already implicitly 'track' this (or,
cannot track it, however you want to view it), so we could at least not
re-issue these requests. It looks like we're already special casing the
receipt of transient responses because of this, so it would have no effect to
reuse the existing responses.
> AbstractReadRepair sends unnecessary data read(s)
> -------------------------------------------------
>
> Key: CASSANDRA-14733
> URL: https://issues.apache.org/jira/browse/CASSANDRA-14733
> Project: Cassandra
> Issue Type: Bug
> Reporter: Benedict
> Priority: Minor
> Labels: Availability, performance
>
> We already have one or more data responses (two in case of 'always'
> speculation, and potentially more if transient replication is enabled, though
> the two do not presently interact).
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
