[
https://issues.apache.org/jira/browse/PHOENIX-6633?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Istvan Toth updated PHOENIX-6633:
---------------------------------
Release Note:
The zkUrl parameter for specifying the Phoenix/HBase ZK quorum has been
deprecated.
The new jdbcUrl parameter has been added, which expected a standard Phoenix
JDBC URL (i.e. jdbc:phoenix...), which is passed to Phoenix without additional
processing.
If neither zkUrl nor jdbcUrl is specified, then Phoenix will fall back to the
default HBase registry determined from hbase-site.xml.
Note that this change only applies to the new DataSourceV2 based "phoenix" data
source.
The old deprecated DataSourceV1 based "org.apache.phoenix.spark" data source
hasn't been updated, and still only works with ZooKeeper based phoenix client
configuration.
> Make ZOOKEEPER_URL optional for Phoenix-Spark connector
> -------------------------------------------------------
>
> Key: PHOENIX-6633
> URL: https://issues.apache.org/jira/browse/PHOENIX-6633
> Project: Phoenix
> Issue Type: Improvement
> Reporter: Istvan Toth
> Assignee: Istvan Toth
> Priority: Minor
>
> We require explicitly specifying the ZK URL for the Phoenix-Spark connector.
> Since having hbase-site.xml on the classpath is required by Phoenix anyway,
> we could use the quorum from hbase-site.xml as default, which would make life
> easier for users.
> Phoenix already handles this, we just need to remove the check, and adopt the
> JDBC URL not to specify the ZK quorum if missing.
> The current zkUrl code also acts as an aritificial limitation on using the
> new zookeeperless Hbase registries from Spark.
> We need to deprecate zkUrl, and add a new jdbcUrl property, which is simply
> passed to Phoenix without further processing.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
