[
https://issues.apache.org/jira/browse/OPENJPA-160?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12475108
]
Patrick Linskey commented on OPENJPA-160:
-----------------------------------------
> 1. Is there a reason why we can't reuse the BrokerImpl objects?
Architecturally, none that I can think of. But we should take care to ensure
that BrokerImpl.free() really does enough work to close up resources
appropriately. Also, I'd love to take a look at some profiling data to see if
we can just optimize creation of brokers instead of adding the complexity of a
pool.
> 2. Assuming we can reuse the objects, where should we put the reuse pool? The
> original implementation created a static map in AbstractBrokerFactory. I'm
> not sure
> that's the best place for it though. BrokerImpl isn't a final class it's
> possible that
> different configurations could use different broker implementations (through
> the
> broker plugin). Maybe we need a new plugin which indicates that class to use
> as
> a Broker pool?
The two options that I see are a Configuration option and a non-static field in
AbstractBrokerFactory. I think that I prefer making it an OpenJPAConfiguration
option, so that it's more easily configurable. Configuration would look like so:
<property name="openjpa.BrokerPool" value="Size=50"/>
If BrokerImpl.free() purges the data passed in to the newBroker() call, then we
should be able to just use a set. In this scenario, the newBroker() code would
then grab something from the set, populate the obtained broker with the data
passed into the newBroker() call, and return it.
If BrokerImpl.free() leaves the broker in a state where the data passed into
newBroker() is relevant, then we should create a key (probably a private inner
class) that includes that data in it, and maintain a map of sets, keyed off of
that data.
All things equal, I'd prefer if we could use a Set (the first case).
> 3. Should we pool the broker instances by default? I think we'll want this to
> be
> configurable, but I'm not sure it needs to be on by default.
We should use the pooling logic, but allow the user to control the pool size.
If this is a performance benefit, then we should choose some reasonable initial
pool size. I have no idea what 'reasonable' is.
> Reuse BrokerImpl objects
> ------------------------
>
> Key: OPENJPA-160
> URL: https://issues.apache.org/jira/browse/OPENJPA-160
> Project: OpenJPA
> Issue Type: Sub-task
> Reporter: Michael Dick
>
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
