Status: Open (View Workflow)
We need a disk full threshold on the master that can stop the queue.
It is all too easy for a Jenkins deployment to fill up disk space on the master and have jobs fail or get stuck in unpredictable ways. The jobs keep coming, making matters worse, and require human time to investigate.
We already have a threshold that will take slaves offline, and it appears there is an advisory that will pop up an alert on the admin console if the master is low on disk space, but we also need an emergency action when storage is critical.
Please add an option for a disk space threshold that will stop the Jenkins queue when storage on the master is critical.
- relates to
JENKINS-7126 Stop All Builds if Master Node is Full
Sorry, I completely forgot about this story. The code in the plugin would benefit from an update to the new Jenkins core baseline, but generally it should work. Have you tried building it from sources brianjmurrell?
Hi! I have created a plugin which solves this issue. https://github.com/AbhyudayaSharma/disk-threshold-plugin . oleg_nenashev Could you please provide a review?
wgc123 of course we can reopen it.
But generally it requires a person who is ready to dedicate hist time for implementing such fix. In this case it can be actually done in a plugin (e.g. by extending http://javadoc.jenkins.io/hudson/model/Queue.QueueDecisionHandler.html ), hence it does not require so much Jenkins developer experience. If you are ready to contribute, I can provide on-demand code reviews.
It looks like this was considered and rejected three years ago in
JENKINS-7126 - can we revisit?
I have not oleg_nenashev. We're not at all set up here to build anything in Java, never mind Jenkins plugins. We're just plain old common end-users of Jenkins at this point.
Although the itchiness of Jenkins and it's plugins, every day makes we wonder if we don't need to start rolling up our sleeves. I just have zero mandate to spend any time doing that at this point.