Random connection refused to nexus snapshot server during deply

[jbalunas]

Hi Paul,

Our RichFaces 3.3.X builds are experiencing a strange problem.  We seem to be getting random connection refused errors like this:

/archive/org.richfaces.ui/modal-panel/3.3.4-20100504.135704-4/modal-panel-3.3.4-SNAPSHOT-sources.jar
[INFO] ------------------------------------------------------------------------
[ERROR] BUILD ERROR
[INFO] ------------------------------------------------------------------------
[INFO] Error installing artifact's metadata: Error while deploying metadata: Failed to create destination WebDAV collection (directory): /nexus/content/repositories/snapshots/org/richfaces/ui/modal-panel/3.3.4-SNAPSHOT

Connection refused

It does not happen at the same point in the build. 

See bottom of:

http://hudson.qa.jboss.com/hudson/view/RF-Dash/job/RichFaces-3.3/566/console

and

http://hudson.qa.jboss.com/hudson/view/RF-Dash/job/RichFaces-3.3/565/console

The strange thing is that the directory from 566 was created successfully and seems ok from quick inspection - https://repository.jboss.org/nexus/content/repositories/snapshots//org/richfaces/ui/modal-panel/3.3.4-SNAPSHOT/

The time was it was created is exactly the time that the build shows.

Does the repo have some limit on connections?  RichFaces does have alot of component, especially the 3.3.X branch.

Any idea's what might be the problem?

Thanks,

Jay

[pgier]

I'm seeing this when downloading artifacts also.  Not sure yet what the cause is though.

[pgier]

Some of the connections from Apache to Nexus were timing out, so the timeout has been increased.  Let me know if the problem continues to happen.

[jbalunas]

Just kicked off a new build to see if that worked.

Thanks Paul

[jbalunas]

Unfortunately it happened again during this build - http://hudson.qa.jboss.com/hudson/job/RichFaces-3.3/569/console

I'll wait to kick it off again until I hear back.

[pgier]

Can you try again with debugging (mvn -X) turned on?

[pgier]

The deployment URL is set incorrectly in the hudson build (dav:https://repository.jboss.org/nexus/content/repositories/snapshots/)

It should be standard http instead of (https://repository.jboss.org/nexus/content/repositories/snapshots/).

This should probably be set in the pom either by extending from the jboss-parent (version 5) or setting the property in the richfaces root pom to a default value since there is currently no default value set for the repository URL.

Prabhat changed the configuration and is restarting the build.  We'll see if this one goes better.

[jbalunas]

What is weird is that it was failing at different places.

[pgier]

That mislead me also.  I didn't think to look at the deployment URL because I would have  though it would fail right way if that was not correct.

Anyway, the builds seem to be working now.

[heiko.braun]

I am still suffering from these problems:

[INFO] snapshot org.jboss.errai:errai-parent:1.1-SNAPSHOT: checking for updates from jboss-public-repository-group
[WARNING] repository metadata for: 'snapshot org.jboss.errai:errai-parent:1.1-SNAPSHOT' could not be retrieved from repository: jboss-public-repository-group due to an error: Error transferring file: Operation timed out
[INFO] Repository 'jboss-public-repository-group' will be blacklisted

Maybe the nexus repo frontend (Noelios-Restlet-Engine) is not configured correctly?

[pgier]

We are seeing some connection timeouts in the Nexus logs, but we haven't found the cause yet.

Can you reproduce the issue with debugging turned on?  (mvn -X)