You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@uima.apache.org by "Marshall Schor (JIRA)" <de...@uima.apache.org> on 2014/12/17 16:17:13 UTC
[jira] [Created] (UIMA-4156) uima-as running a pear scaled out
fails with strange error messages
Marshall Schor created UIMA-4156:
------------------------------------
Summary: uima-as running a pear scaled out fails with strange error messages
Key: UIMA-4156
URL: https://issues.apache.org/jira/browse/UIMA-4156
Project: UIMA
Issue Type: Bug
Components: Async Scaleout
Affects Versions: 2.6.0AS
Reporter: Marshall Schor
Assignee: Jerry Cwiklik
Deploying a Pear as a service, and scaling it out, causes a not-assignable kind of error between a class and an interface, when attempting to initialize the pear for the 2nd scaled-out-instance. Traced this to a conceptual bug: UIMA-AS was passing the same == UimaContext object to the produceAnalysisEngine (in the additionalParams). This object is updated and must be unique per component per scaled-out-instance. Other things in the UimaContext (such as the CAS pool) are needed to be on a per scaled-out-instance, as well. This same issue in the core MultiprocessingAnalysisEngine impl was addressed in UIMA-2078 (see comments)
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)