[
https://issues.apache.org/jira/browse/DRILL-92?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14385660#comment-14385660
]
Yash Sharma commented on DRILL-92:
----------------------------------
[~snazy] + All : On the points mentioned in the review comment -
- Multiple Initial contact points: The patch now handles multiple initial
contact points and passes all the information while connecting to the cluster.
Tested with bringing down certain nodes.
- Session handling : Now the patch uses the cluster instance rather than
session.
- Endpoint affinity & Partition token: The code does consider endpoint affinity
such that - if one of drill endpoint is also a cassandra node it would have an
affinity. The partition token is however for a different purpose. The partition
scheme just ensures that different sub-scans of drill do not fetch the same
range of keys again. It works as a range restriction for each sub-scan. We are
not using it to check where the actual partition data lies.
Tables & Keyspace cache: This is cached not for the functionality for querying
cassandra. Its a schema information which we would need when the user would
like to describe a table or check tables in keyspace etc. We just cache it.
Thanks for the review comments. Please share your thoughts on the new patch.
> Cassandra storage engine
> ------------------------
>
> Key: DRILL-92
> URL: https://issues.apache.org/jira/browse/DRILL-92
> Project: Apache Drill
> Issue Type: New Feature
> Reporter: Steven Phillips
> Assignee: Yash Sharma
> Fix For: Future
>
> Attachments: DRILL-92-Cassandra-Storage.patch, DRILL-92.patch,
> DRILL-CASSANDRA.patch
>
>
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
