jackrabbit-oak-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Chetan Mehrotra (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (OAK-1533) RdbDocumentStore should use a connection pool
Date Tue, 18 Mar 2014 09:01:58 GMT

    [ https://issues.apache.org/jira/browse/OAK-1533?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13938979#comment-13938979
] 

Chetan Mehrotra commented on OAK-1533:
--------------------------------------

bq. which flavour of DS.getConnection do we need? Do we need the variant that takes credentials?

I think maid code should not have any logic around creating DataSource or JDBC Connection.
A DataSource would be provided by the host applications (see below)

bq. what type of dependency the dbcp dependency will be
Preferably it should be a test scope as it would used only in testcode

By default we should remove any logic of creating RDBDocumentStore from builder. Instead for
cases where a RDBDocumentStore is required then it should be specified as part of {{org.apache.jackrabbit.oak.plugins.document.DocumentMK.Builder#setDocumentStore}}.
This is similar to how we handle say custom BlobStore 

So for testcase the code would be creating a DataSource as per test requirement and then create
the RDBDocumentStore from that and pass that to builder

For OSGi case we can have a RDBDocumentStoreFactory which would lookup right DataSource (from
JNDI/Service registry) and then register the DocumentStore with OSGi Service Regitsry and
DocumentNodeStoreService would look up from there

For now I would mark the dependency on dbcp as optional otherwise oak-core bundle would not
start in OSGi env where dbcp bundle is not present





> RdbDocumentStore should use a connection pool
> ---------------------------------------------
>
>                 Key: OAK-1533
>                 URL: https://issues.apache.org/jira/browse/OAK-1533
>             Project: Jackrabbit Oak
>          Issue Type: Task
>          Components: mongomk
>            Reporter: Julian Reschke
>
> The RDB stores should use a connection pool for maximum concurrency.
> In production scenarios, we can probably just use the DataSource API, but we'll need
to figure out what to do in absence of a DataSource (i.e., a test environment).



--
This message was sent by Atlassian JIRA
(v6.2#6252)

Mime
View raw message