Now that the servers activity is calmed down. I've decided to
experiment with this issue again. The result of our build  has
been the same (excepting that we don't have hudson-slave2 server :)
as you have recently announced ):
- hudson-slave1: build succeed
- master: build failed
- fastlane: build failed
In principle, I don't have any problem to be tight to a specific
server, specially now that we have more slots for hudson-server1 :D,
however, I'm curious to know why this may be happening. Is anybody
else experiencing this issue ?
El 15/02/2011 14:16, Denis Roy escribió:
I'm not sure if this is related or not, but the
Master's disk had filled up yesterday with endless Hudson/Java
stack trace spew. I've cleared the log and restarted everything
(yesterday) and all seems well.
On 02/15/2011 06:11 AM, Adolfo Sánchez-Barbudo Herrera wrote:
With the storage problem we have had with the servers, I've
discovered a weird issue building against different servers.
Our buckminster based build  used to be run using
hudson-server1. However, I've discovered that if I run my build
on a different server (hudson-slave2, fastlane and master) the
build fails when trying to execute some Plugin Junit tests with
the following error:
INFO: junit '-DX.display=:12' '--launch' 'org.eclipse.ocl.ecore.tests/launches/org.eclipse.ocl.ecore.tests (Plugin).launch' '--flatXML' '--output' '/opt/users/hudsonbuild/.hudson/jobs/buckminster-mdt-ocl-core-3.1-nightly/workspace//test.results/org.eclipse.ocl.ecore.plugin.tests.xml'
Required plug-in 'org.eclipse.jdt.junit4.runtime' could not be found.
There are other tests which are executed. The difference is that
"org.eclipse.ocl.ecore.tests (Plugin).launch" is a plugin junit tests launch
Before figuring out why this plugin is missed and who needs it,
I'd like to ask you a couple of questions:
- Why could a build success in a server and fail in another one.
Is there any kind of information available ?
- Does anybody face on this/similar issue ?
cross-project-issues-dev mailing list