Community
Participate
Working Groups
As a user it is hard to set up jobs on Hudson, in particular how to tie it to a particular slave or group of slaves, because I don't know what each slave has installed on it, or in what environment jobs are started there. For example, I want to know what versions of Java there are installed on each slave (and where), and other facilities like Ant; network access; build userid; etc.... This enhancement request is to maintain this information somewhere (preferably attached to the hudson admin panels) and find some way of keeping them up-to-date. At minimum it should have the slave's environment variables set up at the start of each job, in the shell that runs jobs. This could be automatically generated and maintained without too much effort. Architecture is indicated on the computer/ pages of the hudson admin, but the information is pitifully meagre.
I haven't been able to locate a way to provide access to that information to regular users via hudson itself. So I updated the wiki.eclipse.org/hudson page with the information you've requested. -M.
Ultimately, we need one set of environment variables and java configuration entries in the slave and master that work across the slaves/master. A job should be able to select say: JDK 5.0 as the JDK to use, and regardless of what server that job runs on, it gets the appropriate JDK. This can be configured on the slaves so they point to particular places, or you allow the Slaves to download and install the necessary JDK if it can't find it. The same needs to be done for Ant, Git, CVS, Subversion, Maven 2 and 3, etc. All of these should be standardized so that moving jobs is just a matter of selecting the correct slave or group to run it against.
