A lot of PVCs left

[kamilkrampa]

This is how our list of PVCs looks in testkube namespace after a couple of days. Is this expected that we're left with so many PVCs?

[vsukhin]

Hey, @kamilkrampa, No, it should not be a case. We use pvc only for container executors, when we need to scape reports after test execution, you need to define artifact parameters in this case. But this pvc is automatically deleted after file scraping. Is this your current flow? Might be,scraper pod was not able to run fro some reason and pvc were not cleaned

Also,would be greate to see jobs and pods in testkube namespace. Especially, when you run or completed the test

[kamilkrampa]

@vsukhin I think we figured out what's the reason. We run testkube tests on spot instances. When an instance is preempted during test execution, a test hangs in running state forever and volumes are not cleaned up.

[vsukhin]

Thank you @kamilkrampa We will keep it there for any other users who face similar issue. Definitely, there are a lot of cases, that we are not able to reproduce without our users 👍 )

[kamilkrampa]

It is also related to this: #3008