[wildfly-dev] Batch Subsystem Model
Cheng Fang
cfang at redhat.com
Fri Sep 13 11:52:17 EDT 2013
On 9/13/13 11:17 AM, James R. Perkins wrote:
> Thanks for all the feedback. If I'm understanding everything right we'll
> end with something like the following for in-memory.
>
> <subsystem xmlns="urn:jboss:domain:batch:1.0">
> <job-repository>
> <in-memory thread-pool="default"/>
> </job-repository>
> <thread-pool name="default">
> <max-threads count="10"/>
> <keepalive-time time="100" unit="milliseconds"/>
> </thread-pool>
> </subsystem>
Why does a job repository need to reference a thread-pool as its
attribute? thread-pool is for job execution and not tied to how jobs
are kept in job repository. By having a top-level <thread-pool>, it
tells this thread-pool configuration is to be used by batch runtime,
regardless in-memory or jdbc job repository.
In your current design, it seems users can configure multiple
<thread-pool> elements and choose one to use. But I don't see this is
really necessary, and still a job repository need not be associated with
a thread-pool.
Cheng
>
> "subsystem" => {
> "batch" => {
> "job-repository" => { "in-memory" => {
> "thread-pool" => "default"
> }},
> "thread-pool" => { "default" => {
> "max-threads" => "10",
> "keepalive-time" => "100",
> "keepalive-unit" => "milliseconds"
> }}
> }
> }
>
> And something like the following for JDBC.
>
> <subsystem xmlns="urn:jboss:domain:batch:1.0">
> <job-repository>
> <jdbc thread-pool="default"
> jndi-name="java:jboss/datasources/ExampleDS"/>
> </job-repository>
> <thread-pool name="default">
> <max-threads count="10"/>
> <keepalive-time time="100" unit="milliseconds"/>
> </thread-pool>
> </subsystem>
>
> "subsystem" => {
> "batch" => {
> "job-repository" => { "jdbc" => {
> "thread-pool" => "default",
> "jndi-name" => "java:jboss/datasources/ExampleDS"
> }},
> "thread-pool" => { "default" => {
> "max-threads" => "10",
> "keepalive-time" => "100",
> "keepalive-unit" => "milliseconds"
> }}
> }
> }
>
> I personally prefer the <job-repository><in-memory/></job-repository>
> over the <in-memory-job-repository/>, but I don't care all that much. If
> the consensus is the later, that's fine.
>
> Also do we want the thread pool to be called batch-thread-pool or just
> simply thread-pool?
>
> Thanks for all the feedback. I just wanted to get it right since what we
> decide we're stuck with.
>
More information about the wildfly-dev
mailing list