[ 
https://issues.apache.org/jira/browse/IGNITE-1179?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14653189#comment-14653189
 ] 

Darren Edmonds commented on IGNITE-1179:
----------------------------------------

I don't believe the futures are hanging, I don't think they're being executed.

I have attached the code that submits the job to the workers and the future 
logging that should appear when it's complete (which according to the 
worker.log, it does?).
{code:title=SchedulerService.java|borderStyle=solid}
    ...    
    LOG.warn("DEBUG: JOB [{}] Job submitted to worker node.", 
coreJob.getKey().toString());
    //IgniteCompute task = 
grid.compute(grid.cluster().forRemotes()).withAsync();
    IgniteCompute task = 
grid.compute(grid.cluster().forRemotes().forAttribute("ROLE", 
"WORKER")).withAsync();
    task.affinityCall("ASSET", coreJob.getKey(), nodeJob);
    //task.call(nodeJob);
    IgniteFuture<NodeJobResult> fut = task.future();
    //fut.listen(listenerFactory.getListener(coreJob));
    fut.listen(new IgniteInClosure<IgniteFuture<NodeJobResult>>() {
        @Override
        public void apply(IgniteFuture<NodeJobResult> 
nodeJobResultIgniteFuture) {
            LOG.warn("DEBUG: JOB [{}] Future called.", 
coreJob.getKey().toString());
            
listenerFactory.getListener(coreJob).apply(nodeJobResultIgniteFuture);
        }
    });
    ...
{code}

> Futures stop working, unknown cause.
> ------------------------------------
>
>                 Key: IGNITE-1179
>                 URL: https://issues.apache.org/jira/browse/IGNITE-1179
>             Project: Ignite
>          Issue Type: Bug
>          Components: compute
>            Reporter: Darren Edmonds
>            Assignee: Andrey Gura
>         Attachments: ApplicationLogsAndConfig_1.3.2.7z
>
>
> Running against Apache Ignite version: 
> ignite-core (1.3.2), ignite-spring (1.3.2), ignite-indexing (1.3.2) and 
> ignite-slf4j (1.3.2).
> Originally built application against a Grid Gain version before moving to 
> Apache Ignite.
> Web Application feeds jobs to the worker node(s) via it's own internal queue 
> (client only Ignite).  After a few thousand jobs have successfully been 
> processed, the worker and server just stop with jobs still waiting in the web 
> server queue.  It appears that the worker node is running OK, and logs will 
> show every job is successfully completing the render but the future on the 
> server is not being triggered - loss of connection to node?
> I have attached the requested information to the ticket.  It's highly likely 
> I've done something wrong rather than this being a bug with Ignite - any 
> advice would be appreciated,
> Attachment contains:
> - Server stack
> - Worker node stack
> - Server log file (simple debug output, no output observed from Ignite set at 
> INFO level nor any exceptions)
> - Worker log file (same setup as server log file)
> - Ignite configuration files for both server and worker.
> - Launch parameters for the java command line and tomcat7 server.
> Original forum post:
> http://apache-ignite-users.70518.x6.nabble.com/Futures-not-being-triggered-td773.html



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to