[
https://issues.apache.org/jira/browse/UIMA-4964?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15344940#comment-15344940
]
Lou DeGenaro commented on UIMA-4964:
------------------------------------
1. update default ducc.properties to include:
> ducc.ws.state.update.endpoint=ducc.ws.state
> ducc.ws.state.update.endpoint.type=topic
> ducc.ws.state.publish.rate=5000
2. update DUCC Book for the above additions
3. update Common Configuration for the above additions
4. add new transport Ducc Event: WEBSERVER_STATE
5. update webserver
> daemons data and broker helper to determine if self publications are
stale
> webserver component & event listener to receive and process self
publications
> create webserverstate, webserverstateduccevent, webserverstateprocessor
to send and receive self publications
> update handlers (classic & json) to accurately reflect daemons page
status of broker and other daemons with respect to broker status
> DUCC Web Server (WS) should employ more reliable aliveness detection for
> database & broker
> ------------------------------------------------------------------------------------------
>
> Key: UIMA-4964
> URL: https://issues.apache.org/jira/browse/UIMA-4964
> Project: UIMA
> Issue Type: Bug
> Components: DUCC
> Reporter: Lou DeGenaro
> Assignee: Lou DeGenaro
> Fix For: 2.1.0-Ducc
>
>
> Presently, the Web Server (WS) determines up/down for the database (DB) via
> JMX. A better test would be to perform an actual DB query or perhaps even
> confirm that actual DB updates succeed.
> Likewise, a better test of broker aliveness would be to send actual messages
> from WS to WS via broker to insure successful round trip.
> Note: such changes may be problematic for "shadow" WS when not running as
> actual "ducc" user.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
