[jira] [Updated] (PHOENIX-2359) Configuration for PQS to use Protobuf serialization instead of JSON

["Josh Elser (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/PHOENIX-2359?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Josh Elser updated PHOENIX-2359:
--------------------------------
    Attachment: PHOENIX-2359.002.patch

.002 is a slight improvement on sqlline-thin.py. Using {{hbase}} on the {{$PATH}}, it invokes HBaseConfTool to extract the value of {{phoenix.queryserver.serialization}} from hbase-site.xml and automatically updates the JDBC URL.

If {{hbase}} isn't on the PATH, the value isn't defined (or there is some problem invoking HBaseConfTool), it defaults to JSON.

> Configuration for PQS to use Protobuf serialization instead of JSON
> -------------------------------------------------------------------
>
>                 Key: PHOENIX-2359
>                 URL: https://issues.apache.org/jira/browse/PHOENIX-2359
>             Project: Phoenix
>          Issue Type: Improvement
>            Reporter: Josh Elser
>            Assignee: Josh Elser
>            Priority: Trivial
>             Fix For: 4.7.0
>
>         Attachments: PHOENIX-2359.001.patch, PHOENIX-2359.002.patch
>
>
> Calcite-1.5.0 is soon to drop. Included in that release is some work to use protocol buffers for the data transported over HTTP between client and server instead of JSON.
> I made some changes in Avatica which hopefully make Phoenix's use of the Avatica server easier, in addition to choosing the serialization.
> Let me stage the patch I've been using to test this. We can't apply it until we're confident that moving to the newer version of Calcite is safe to do.



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