[
https://issues.apache.org/jira/browse/AIRAVATA-1620?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14350902#comment-14350902
]
ASF subversion and git services commented on AIRAVATA-1620:
-----------------------------------------------------------
Commit 774b092d31e41919b39ca3ce9f1edd5af1c30669 in airavata's branch
refs/heads/new-workflow-design-rabbitmq from [~shameera]
[ https://git-wip-us.apache.org/repos/asf?p=airavata.git;h=774b092 ]
Fixed AIRAVATA-1620.
> Optimize thread usage in Workflow Enactment Service.
> ----------------------------------------------------
>
> Key: AIRAVATA-1620
> URL: https://issues.apache.org/jira/browse/AIRAVATA-1620
> Project: Airavata
> Issue Type: Improvement
> Reporter: Shameera Rathnayaka
> Assignee: Shameera Rathnayaka
>
> Currently new workflow model allocates a thread from the thread pool to run
> the workflow and thread comes back to the thread pool after finishing the all
> workflow nodes executed in that workflow. As the nature of the workflows runs
> on real case, this workflows may take days to complete. In this case, it is
> not idea to allocate a thread until it finishes, most of the time thread is
> in sleep state and only need to run once the task events trigger for the
> submitted task.
> We can optimize the thread utilization by allowing thread to come back to the
> thread pool till the next task output and status event trigger. Workflow
> enactment service keep listening to the all task events and invoke correct
> workflow and hand over the event.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
