[
https://issues.apache.org/jira/browse/IGNITE-25380?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Pavel Pereslegin updated IGNITE-25380:
--------------------------------------
Description:
Currently, the JDBC driver only connects to one cluster node, which deprives it
of the ability to optimize interaction with the cluster.
Support for multiple simultaneous connections to different cluster nodes will
allow the JDBC driver to support the following functionality:
* Load balancing
* Partition awareness
* Automatic connection retry
Under the hood the JDBC driver uses a java thin client to establish connection
to the cluster node. The driver already allows the user to specify multiple
addresses for connection. However, to execute queries, the driver uses its own
requests and operation types, and this implementation is designed to work with
only one node.
At the same time, the SQL API implementation over a Thin Client already
supports work with multiple nodes, as well as load balancing and partition
aware query execution. Therefore, it is proposed to migrate the JDBC operations
to use the SQL API over Thin Client.
was:
Currently, the JDBC driver only connects to one cluster node, which deprives it
of the ability to optimize interaction with the cluster.
Support for multiple simultaneous connections to different cluster nodes will
allow the JDBC driver to support the following functionality:
* Load balancing
* Partition awareness
* Automatic connection retry
Currently under the hood the JDBC driver uses a java thin client to establish
connection to the cluster node. The driver already allows the user to specify
multiple addresses for connection. However, to execute queries, the driver uses
its own requests and operation types, and this implementation is designed to
work with only one node.
At the same time, the SQL API implementation over a Thin Client already
supports work with multiple nodes, as well as load balancing and partition
aware query execution. Therefore, it is proposed to migrate the JDBC operations
to use the SQL API over Thin Client.
> Support multiple endpoints in JDBC connection
> ----------------------------------------------
>
> Key: IGNITE-25380
> URL: https://issues.apache.org/jira/browse/IGNITE-25380
> Project: Ignite
> Issue Type: Epic
> Components: jdbc ai3
> Reporter: Iurii Gerzhedovich
> Priority: Major
> Labels: ignite-3
>
> Currently, the JDBC driver only connects to one cluster node, which deprives
> it of the ability to optimize interaction with the cluster.
> Support for multiple simultaneous connections to different cluster nodes will
> allow the JDBC driver to support the following functionality:
> * Load balancing
> * Partition awareness
> * Automatic connection retry
> Under the hood the JDBC driver uses a java thin client to establish
> connection to the cluster node. The driver already allows the user to specify
> multiple addresses for connection. However, to execute queries, the driver
> uses its own requests and operation types, and this implementation is
> designed to work with only one node.
> At the same time, the SQL API implementation over a Thin Client already
> supports work with multiple nodes, as well as load balancing and partition
> aware query execution. Therefore, it is proposed to migrate the JDBC
> operations to use the SQL API over Thin Client.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
