Hi All, In our team we are trying to treat jenkins configuration as source code.
We source control it, check it out and run it against local resources while we're maintaining it and then deploy it to the actual jenkins server which uses the actual resources. We therefore have many jenkins installations, running the same jobs, but only one of the installations is the real one, and the others are development installations where we work on actual jobs. (Similar to how you run an applications locally in the debugger against local resources while you are writing the code and push it out afterwards) We are running into a few difficulties with this workflow but we have OK workarounds for a lot of them. One issue that is bugging me is that of the node configuration. Each of us will have a list of nodes that look the same, but due to machine differences some of those things might be different, eg, tool locations, folder locations, workspace location etc. I tried to use "$WORKSPACE_DIR" (ie environment variable) in the 'remote fs root' setting, but it does not get resolved on the node. I see two solutions to this problem, but there might be more and I would appreciate some help/advice here: - We can change the jenkins node runner to resolve environment variables in the configuration settings before jobs run, then we can abstract away the differences between individual jenkins installations this way - We can change the jenkins configuration code to store the nodes config in a separate file and then not check that into source control. Are there other avenues that I'm missing here? Any other ways of looking at this problem that might make sense? Kind regards, Pieter -- You received this message because you are subscribed to the Google Groups "Jenkins Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to jenkinsci-users+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/groups/opt_out.
