[
https://issues.apache.org/jira/browse/GORA-214?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13601360#comment-13601360
]
Roland von Herget commented on GORA-214:
----------------------------------------
Hi Renato,
1) I do not know. Sorry :/
2) In my patch the executeMutator() is called for every key we addOrUpdate (see
the one line code & two lines comment patch to CassandraStore)
So, you will find real improvements only if we update more than one column for
one key.
As far as I read about this, grouping together multiple mutations for one key,
will reduce additional overhead inside Cassandra. In contrast to grouping
together mutations for multiple keys will only save time for e.g. network
latency, but nothing else.
I'm sorry, I can't cite anything here, I can't remember where I read about it.
--Roland
> use batched Mutations available in Hector Mutator
> -------------------------------------------------
>
> Key: GORA-214
> URL: https://issues.apache.org/jira/browse/GORA-214
> Project: Apache Gora
> Issue Type: Improvement
> Components: storage-cassandra
> Reporter: Roland von Herget
> Assignee: Roland von Herget
> Fix For: 0.3
>
> Attachments: GORA-214-prove-of-concept-0.2.patch,
> GORA-214-trunk-concept.patch
>
>
> Cassandra & Hector offer a lot of possibilities to batch mutations together.
> This means Hector will not send data out to Cassandra until it receives the
> command to do so [execute()].
> Really useful is this feature if you have to addOrUpdate two or more columns
> for one key (in one cf), it will then limit load on Cassandra, too.
> A small prove of concept patch against 0.2 is attached.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
