Hi Karen,
As a brand new developer, I'm afraid I can't be much help on this. I
still don't have much experience interpreting workflows, especially
their failures, so will have to defer to someone more knowledgeable.
John King
On 9/26/12 5:53 AM, Karen Dolan wrote:
John and Matterhorn workflow fans,
The fix John submitted publishes dublincore to new distributions that
go through a full workflow. We are trying to adapt the fix with a
short workflow applied to episodes, see below, to publish dublincore
to already published distributions. The short workflow tags
dublincore/episode, dublincore/series and the mpeg-7/text flavors for
publish.
The problem is that a bad mpeg-7 file is being sent by the workflow.
The Admin interface, Inspect Workflow -> Mediapackage, has the
mpeg-7/text location to a good file located on the Episode server.
But, that good file is not being accessed during the short workflow.
The path to an empty file on the Admin server is accessed instead.
My hunch is that a temporary mpeg-7 was created on the Admin during
it's first workflow. Somehow, the file was not archived but the link was.
Course of action...
1) Figure out how admin knows about a good link, but the bad link
takes over in the workflow.
2) Figure out how the good link can be tagged and sent in publish step.
3) If it's really necessary to have the mpeg-7 on the Admin, ensure
the mpeg-7 file is archived on the Admin for future workflows. It
doesn't help the current problem.
I am grateful for any insights, suggestions, or alternative
approaches before I dig deeper.
Best Regards,
Karen
More Background
In the Matterhorn Admin's Inspect Workflow -> Mediapackage there are
good links to the the appropriate xml files. The mpeg-7/text link is
to the Engage server, not the Admin server where the dublincore files
are.
During the workflow's publish step, the Engage server chokes with a
bad link to the mpeg-7 file. Engage can not parse the empty xml in the
link sent, so it logs the error and leaves the workflow hanging. The
link sent was to an empty file on the Admin server, not the good link
to the mpeg-7 file on the Engage server.
"2012-09-25 16:43:50 WARN (StaticMetadataServiceDublinCoreImpl:293) -
Unable to load metadata from catalog
'http://admin...edu/episode/archive/mediapackage/12a663c2-78c1-4ace-bc95-f60fec00ee4b/49037340-33e9-43f8-8a9d-7dba29d0cbef/0'
2012-09-25 16:43:50 DEBUG (SolrIndexManager:383) - Looking for mpeg-7
catalogs containing segment texts ...."
--------------------------------------------
1) Admin Tool Inspect Workflow -> Mediapackage links
-------------------------------------
dublincore/episode
http://admin...edu/files/mediapackage/12a663c2-78c1-4ace-bc95-f60fec00ee4b/49037340-33e9-43f8-8a9d-7dba29d0cbef/dublincore.xml
dublincore/series
http://admin...edu/files/mediapackage/12a663c2-78c1-4ace-bc95-f60fec00ee4b/43807a83-849c-45c7-8c22-02bbd2e693f5/dublincore.xml
mpeg-7/text
http://engage...edu/static/12a663c2-78c1-4ace-bc95-f60fec00ee4b/catalog-4/slidetext.xml
[Observe, mpeg-7 xml link is to the engage server, not the admin.]
---------------------------------------------------------
2) Workflow Applied to Episodes to try get their dublin core over to
search:
-----------------------------------------------------------
<?xml version="1.0" encoding="UTF-8"?>
<definition xmlns="http://workflow.opencastproject.org";>
<!-- Description -->
<id>DCE-fix-dublin-core-try-3</id>
<title>DCE: Workaround to add dublin core to already published</title>
<description>
Tag source flavors dublincore/* and mpeg-7 for target publish.
Publish to search index.
</description>
<!-- Operations -->
<operations>
<operation id="tag" description="Tagging dublin core catalogs for
publishing">
<configurations>
<configuration
key="source-flavors">mpeg-7/text,dublincore/*</configuration>
<configuration key="target-tags">publish</configuration>
</configurations>
</operation>
<!-- Publish to engage search index -->
<operation id="publish" fail-on-error="true"
exception-handler-workflow="error" description="Published">
<configurations>
<configuration key="source-tags">publish</configuration>
</configurations>
</operation>
</operations>
</definition>
-------------------------------------------
3) Engage Error stacktrace in matterhorn.log "
----------------------------------
2012-09-25 16:43:50 DEBUG (WorkspaceImpl:249) -
http://admin1.mh-dev.harvard.edu/episode/archive/mediapackage/12a663c2-78c1-4ace-bc9
5-f60fec00ee4b/49037340-33e9-43f8-8a9d-7dba29d0cbef/0 has not been
modified.
2012-09-25 16:43:50 WARN (DublinCoreCatalogImpl$DublinCoreParser:843)
- Fatal error parsing dublincore catalog: Premature end of file.
2012-09-25 16:43:50 WARN (StaticMetadataServiceDublinCoreImpl:293) -
Unable to load metadata from catalog
'http://admin...edu/episode/archive/mediapackage/12a663c2-78c1-4ace-bc95-f60fec00ee4b/49037340-33e9-43f8-8a9d-7dba29d0cbef/0'
2012-09-25 16:43:50 DEBUG (SolrIndexManager:383) - Looking for mpeg-7
catalogs containing segment texts
2012-09-25 16:43:50 DEBUG (WorkspaceImpl:249) -
http://admin...edu/episode/archive/mediapackage/12a663c2-78c1-4ace-bc9
5-f60fec00ee4b/catalog-4/0 has not been modified.
2012-09-25 16:43:50 WARN (Mpeg7Parser:397) - Fatal error while
parsing mpeg-7 catalog: Premature end of file.
2012-09-25 16:43:50 WARN (SearchRestService:107) -
org.apache.solr.client.solrj.SolrServerException:
java.lang.IllegalStateExceptio
n: Unable to load mpeg-7 catalog data:Premature end of file.
org.opencastproject.search.api.SearchException:
org.apache.solr.client.solrj.SolrServerException:
java.lang.IllegalStateException: U
nable to load mpeg-7 catalog data:Premature end of file.
at
org.opencastproject.search.impl.SearchServiceImpl.add(SearchServiceImpl.java:374)
at
org.opencastproject.search.endpoint.SearchRestService.add(SearchRestService.java:104)
at sun.reflect.GeneratedMethodAccessor195.invoke(Unknown Source)
at
sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.lang.reflect.Method.invoke(Method.java:616)
at
org.apache.cxf.service.invoker.AbstractInvoker.performInvocation(AbstractInvoker.java:173)
at
org.apache.cxf.service.invoker.AbstractInvoker.invoke(AbstractInvoker.java:89)
...
_______________________________________________
Matterhorn mailing list
[email protected]
http://lists.opencastproject.org/mailman/listinfo/matterhorn
To unsubscribe please email
[email protected]
_______________________________________________
--
John King
Applications Programmer
Learning Systems Group
Educational Technology Services
9 Dwinelle Hall - Mail
117 Dwinelle Hall - Office
University of California
Berkeley, CA 94720-2535
Phone: 510-529-5074
Email: [email protected]
_______________________________________________
Matterhorn mailing list
[email protected]
http://lists.opencastproject.org/mailman/listinfo/matterhorn
To unsubscribe please email
[email protected]
_______________________________________________
