[
https://issues.apache.org/jira/browse/BROOKLYN-197?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15022207#comment-15022207
]
Svetoslav Neykov commented on BROOKLYN-197:
-------------------------------------------
Jetty is doing the right thing - intercepting the dtd lookups of the xml parser
(for web.xml) and redirecting them to jar locations (see
https://github.com/eclipse/jetty.project/blob/master/jetty-webapp/src/main/java/org/eclipse/jetty/webapp/WebDescriptor.java#L141).
The problem is that the files can't be found on the classpath. Adding a
dependency on jetty schemas
(http://mvnrepository.com/artifact/org.eclipse.jetty.toolchain/jetty-schemas)
should be good enough. Note that pre-jetty 9 we depended on a jetty version of
the servlet api which included the schemas
(https://github.com/apache/incubator-brooklyn/pull/997/files#diff-9983db05471eb0c39f1c234505c51ac0L124)
> Fails to launch unless Internet access is available
> ---------------------------------------------------
>
> Key: BROOKLYN-197
> URL: https://issues.apache.org/jira/browse/BROOKLYN-197
> Project: Brooklyn
> Issue Type: Bug
> Affects Versions: 0.9.0
> Reporter: Richard Downer
>
> Brooklyn 0.9.0-SNAPSHOT requires a connection to the public Internet to work.
> This is a regression from Brooklyn 0.8.0.
> This is an issue when Brooklyn is launched into secured server environments.
> It is common in many corporations that production servers have restricted
> egress access to the public Internet. Therefore this is a major issue.
> The issue appears to be that Jetty is using Xerces to parse XML files; Xerces
> is trying to resolve the DTDs by downloading them from the URL in the XML
> header. Therefore, it will be trying to access http://java.sun.com. When this
> operation fails, the root webapp is not deployed, and the Brooklyn web
> console (and probably the REST API too) returns 503 Service Unavailable.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
