add request context to org.eclipse.jetty.io.EofException to aid determining the root cause

An extremely common scenario is HTTP clients dropping the connection to Nexus, before Nexus completes writing the response.

This typically results in many WARN messages printed in the logs such as

jvm 1 | 2014-04-08 07:12:18 WARN [tp693255375-837] - org.sonatype.nexus.web.content.NexusContentServlet - org.eclipse.jetty.io.EofException, caused by: java.io.IOException: Connection reset by peer
jvm 1 | 2014-04-08 07:12:20 WARN [p693255375-1433] - org.sonatype.nexus.web.content.NexusContentServlet - org.eclipse.jetty.io.EofException, caused by: java.io.IOException: Broken pipe

For the most part this can be safely ignored - but still logs fill with WARN messages lacking context which is concerning to support and customers.

Some not so obvious causes:

• Maven lightweight Wagon bugs ( https://jira.codehaus.org/browse/WAGON-384 )
• sometimes corporate firewalls and virus scan the requested file and force drop the connection
• custom written HTTP clients accessing Nexus have been known to trigger this

This is handled in 2.8 here:

https://github.com/sonatype/nexus-oss/blob/nexus-2.8.x/plugins/basic/nexus-content-plugin/src/main/java/org/sonatype/nexus/content/internal/ContentServlet.java#L314-314

This issue proposes to improve the log message printed to include

• UA
• HTTP method
• request path
• request host/IP address

With this information, we can more accurately and quickly deduce if the EofException is something that should be investigated further by a Nexus Administrator.