Hi guys,
as we were having some trouble successfully running NUnit from Continua CI which seems not to be Continua CI related I activate the dedicated cleanup rule for our build configuration and disabled cleanup for failed builds. This configuration has multiple steps that are assigned to different hosts by using strict agent requirements. I would have expected that the cleanup is performed after every step and only the workspace for the failed steps are ignored. It seems that all workspaces on all agents are not cleaned up if a single step has failed. Is this the intended behavior?
I’d like to be able to clean up the workspace after every successful step and just keep the workspace of the failed steps to analyze the error condition.
Kind regards
Kay Zumbusch
Hi Kay,
The agent workspace cleanup happens at the end of the build and the cleanup rules are based on the status of the build rather than the status of each stage. If the build status matches one of the selected status options then a message is sent to all agents to cleanup up their workspace for that build, otherwise no message is sent.
We could add an option to change this behaviour, but generally we think that it would be more useful to keep the workspace of every stage when a build fails. A failure in one stage could be dependant on some files which were not generated in a previous stage.
Is disk space the reason for wanting to clean up the previous stages? Note that the agent workspace is always cleaned up when the Server Cleanup Policy runs irrespective of the Agent Workspace Cleanup Policy.
Hi Dave,
that sounds reasonable. Our agents are always cleaned up after a build even for defective builds. We just change the agent cleanup policies if we need to debug a failing step. And in this case we would have to cleanup all non defective agents manually to clear the blocked disk space. It is a rare case for us but still a bit annoying.
Kind regards
Kay Zumbusch