I have installed the Join Trigger plugin on my Hudson (version 1.336) which will allow job X to run job Y after all downstream jobs from job X have completed. Once in a while, however, I don't want job X to join any other jobs.
One way of implementing this, is to use Groovy to identify and delete job X from the runtime model, and then use the hudson-cli to create-job, passing in a config.xml for job X that does NOT contain the Join trigger configuration:
Even though I see job X's configuration updated correctly on the Hudson GUI, the runtime model doesn't get updated with the new configuration. job X still joins with it's jobs.
The only way around this is to explicitly reload all configurations from disk (which defeats the purpose of runtime adjustments).