I'm configuring an environment with a jenkins master that uses kubernetes to instantiate slaves. Everyting works fine except that the workspace is not available when the slave is killed by the kubernetes (slave offline). I have tried several solutions (using an NFS, copying back the workspace on the master with the copy-to-slave plugin, ... aso) .... but it doesn't work ... Jenkins don't want to display the workspace if the node is offline (even if the workspace is at the right location, but on the master). Do you have a solution ?
Carlos Sanchez Could you please help with this issue, or how we can fix it?
Olivier Lambert Could you please let me know if you were able to solve the issue?
We do have a workaround which is not that nice.. Once the build is done, we have to change the 'where the project is run ' to master. Then the workspace is shown.
Raghu Pallikonda We also find this workaround. It works fine .... but it's just a workaround ... We think a better solution should be implemented directlry in the core of Jenkins. 1. Check if the "where the project is run" is online 2. IF YES >> display this workspace 3. IF NO >> check if it's available on the master 4. IF YES, display it 5. IF NO, display error message
Olivier Lambert, looks like this PR was never approved. How did you get around this issue? Carlos Sanchez We have the same exact issue. Looks like the PR was never approved. Can we submit a new PR for this?
Carlos Sanchez, using copy-to-slave plugin is a terrible idea. Because that plugin isn't maintained and is even taken off of the official Jenkins plugin list due to unfixed security problems. Being able to view the workspace after the slave node is destroyed is quite important and would be very useful to have that feature built in.
I am also facing the issue where we would like to see the job's workspace after the pod terminated . Our agents are attached to a PVC ( persistence volume ) and can still see the workspace even after the pod terminates but cannot access through the job bcz no pod exist.
please let me know if you found ways to address this issue