You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@uima.apache.org by "Jerry Cwiklik (JIRA)" <de...@uima.apache.org> on 2016/10/10 17:31:20 UTC
[jira] [Closed] (UIMA-5142) DUCC: JP scaleout not derived from
process_pipeline_count for DD jobs
[ https://issues.apache.org/jira/browse/UIMA-5142?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jerry Cwiklik closed UIMA-5142.
-------------------------------
Resolution: Fixed
Modified to use -Dducc.deploy.JpThreadCount=X for scaleout when initializing a JP. The X determines number of threads the JP will use to fetch WIs from a JD.
> DUCC: JP scaleout not derived from process_pipeline_count for DD jobs
> ---------------------------------------------------------------------
>
> Key: UIMA-5142
> URL: https://issues.apache.org/jira/browse/UIMA-5142
> Project: UIMA
> Issue Type: Bug
> Components: DUCC
> Reporter: Jerry Cwiklik
> Assignee: Jerry Cwiklik
> Fix For: 2.1.0-Ducc
>
>
> For DD based jobs the scaleout is not based on process_pipeline_count specified in job's spec. Instead, the code tries to fetch the scaleout from a DD which is not right.
> When a job is submitted, the OR creates a command line to launch a JP to run the job. The OR takes the value of process_pipeline_count parameter and creates a System property from it as follows:
> arg[4]: -Dducc.deploy.JpThreadCount=2
> All the JP needs to do is to fetch the property value and use it to create matching number of process threads that will request WIs from a JD.
>
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)