When archiving artifacts, if the host machine is out of disk space the archive artifacts feature does not fail the build. In my case this means any downstream builds dependent on these artifacts kick off and fail due to the artifacts not existing. Example:

23:52:34 Archiving artifacts
23:53:36 ERROR: Failed to archive artifacts: coreteststag, viscob/installer/*/.exe, viscob/installer/*/.msi, viscob/vscore/testdata.zip, viscob/cobdir/Release/UpdateSite/*, viscob/testing/debugharness/, viscob/eclipseide/testware/*
23:53:36 hudson.util.IOException2: hudson.util.IOException2: Failed to extract c:\jenkins\workspace\R5_Win_Overnight/coreteststag, viscob/installer/*/.exe, viscob/installer/*/.msi, viscob/vscore/testdata.zip, viscob/cobdir/Release/UpdateSite/*, viscob/testing/debugharness/, viscob/eclipseide/testware/*
23:53:36 at hudson.FilePath.readFromTar(FilePath.java:1662)
23:53:36 at hudson.FilePath.copyRecursiveTo(FilePath.java:1580)
23:53:36 at hudson.tasks.ArtifactArchiver.perform(ArtifactArchiver.java:116)
23:53:36 at hudson.tasks.BuildStepMonitor$1.perform(BuildStepMonitor.java:19)
23:53:36 at hudson.model.AbstractBuild$AbstractRunner.perform(AbstractBuild.java:682)
23:53:36 at hudson.model.AbstractBuild$AbstractRunner.performAllBuildSteps(AbstractBuild.java:657)
23:53:36 at hudson.model.AbstractBuild$AbstractRunner.performAllBuildSteps(AbstractBuild.java:635)
23:53:36 at hudson.model.Build$RunnerImpl.post2(Build.java:162)
23:53:36 at hudson.model.AbstractBuild$AbstractRunner.post(AbstractBuild.java:604)
23:53:36 at hudson.model.Run.run(Run.java:1401)
23:53:36 at hudson.model.FreeStyleBuild.run(FreeStyleBuild.java:46)
23:53:36 at hudson.model.ResourceController.execute(ResourceController.java:88)
23:53:36 at hudson.model.Executor.run(Executor.java:230)
23:53:36 Caused by: java.io.IOException: There is not enough space on the disk
23:53:36 at java.io.FileOutputStream.writeBytes(Native Method)

It would be useful if such an error could fail the build - this would stop other downstram builds from starting and causing multiple build failures. It would also make it easier to see where the original problem occurred.