You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Vladimir Ermakov (Jira)" <ji...@apache.org> on 2022/06/10 10:29:00 UTC

[jira] [Assigned] (IGNITE-16987) JDBC driver for 3.0: connection channel usage refactoring

     [ https://issues.apache.org/jira/browse/IGNITE-16987?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Vladimir Ermakov reassigned IGNITE-16987:
-----------------------------------------

    Assignee: Vladimir Ermakov

> JDBC driver for 3.0: connection channel usage refactoring
> ---------------------------------------------------------
>
>                 Key: IGNITE-16987
>                 URL: https://issues.apache.org/jira/browse/IGNITE-16987
>             Project: Ignite
>          Issue Type: Improvement
>          Components: jdbc, sql
>            Reporter: Vladimir Ermakov
>            Assignee: Vladimir Ermakov
>            Priority: Major
>              Labels: ignite-3
>
> 1 part:
> In 3.0 JDBC all queries go through ReliableChannel, which can do reconnects, retries, etc.
> This is wrong for things like SQL_NEXT, SQL_QUERY_META, SQL_CURSOR_CLOSE, and so on, which should be executed strictly on the same connection. To do this, there's PayloadInputChannel.clientChannel , which you can keep when a query/cursor is initiated.
>  
> 2 part:
> We have to close resources on the server after client is disconnected. To do this, there is a ClientResourceRegistry , it is cleared when you disconnect. If that doesn't work, see ClientInboundMessageHandler#channelInactive



--
This message was sent by Atlassian Jira
(v8.20.7#820007)