cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Carl Yeksigian (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (CASSANDRA-10813) Add 'idempotent' flag to 'PREPARED' response
Date Thu, 03 Dec 2015 18:38:11 GMT

     [ https://issues.apache.org/jira/browse/CASSANDRA-10813?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Carl Yeksigian updated CASSANDRA-10813:
---------------------------------------
    Issue Type: Sub-task  (was: Improvement)
        Parent: CASSANDRA-9362

> Add 'idempotent' flag to 'PREPARED' response
> --------------------------------------------
>
>                 Key: CASSANDRA-10813
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-10813
>             Project: Cassandra
>          Issue Type: Sub-task
>            Reporter: Andy Tolbert
>            Priority: Minor
>
> I'm not sure how attainable this is, but it would be very convenient to know whether
or not a statement that is prepared is idempotent.  If at all possible it would be nice to
add a bit to flags that communicates this in a future protocol version.
> This would be useful for client library implementors for making decisions based on whether
or not a query is idempotent.  For example, if a connection fails while a request is in flight,
you can't be sure whether or not a write mutation was applied or not.  If the write query
was idempotent and client timestamps were used, you can feel comfortable simply retrying the
request.  The datastax drivers have the capability of specifying whether or not a query is
idempotent manually, but it would be nice if C* could tell us whether or not a query is idempotent
at prepare time (if it all possible) instead of depending on the user to tell us.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message