[jira] [Commented] (CASSANDRA-12663) Allow per DC segregation, grant user to create different indices per datacenter on tables

["Bhuvan Rawal (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CASSANDRA-12663?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15536559#comment-15536559 ] 

Bhuvan Rawal commented on CASSANDRA-12663:
------------------------------------------

From a users point of view a way of  index creation could be:
CREATE INDEX user_state_index ON keyspace.users (state) WITH replication ['dc1','dc2'];

If secondary index is being queried on a DC where index does not exist the coordinator can proxy it to a healthy DC where index lies. Similar to how a coordinator routes query if a Keyspace does not belong to its DC.

> Allow per DC segregation, grant user to create different indices per datacenter on tables
> -----------------------------------------------------------------------------------------
>
>                 Key: CASSANDRA-12663
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-12663
>             Project: Cassandra
>          Issue Type: New Feature
>          Components: Core
>            Reporter: Bhuvan Rawal
>
> For analytics & auditing purposes it becomes essential to serve different access patterns than that modeled from a partition key fetch perspective, although a limited reads are needed by users but if enabled cluster wide it will require index write for every row written on that table on every single node on every DC even the one which may be serving read operations. A user may not want to have indices built on Transactional DC on every write, that computation and disk utilization may not be useful as the Analytics may possibly be performed on other DC.
> It will be a plus to have analytics / auditing workload built inside Cassandra itself using native secondary indices / SASI indices / Stratio by creating indices for a specific datacenter and not having to ship off data to other index stores like Elasticsearch through application.



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