#132 CI test ran out of disk space
Closed by bgoncalv. Opened by tstellar.

A recent CI test of mine ran out of disk space:
https://jenkins-continuous-infra.apps.ci.centos.org/blue/organizations/jenkins/fedora-f33-build-pipeline/detail/fedora-f33-build-pipeline/28/pipeline/

Is there any way to request additional disk space? What is the default?


I checked with our team, and these default to 14GB of space; sadly, there's no way to define more space as needed.

How much space did you need for your tests; do you know?

We could probably increase the filesystem in the qcow2, but this will be valid for every build that's why we'd like to know how much free space your test requires.

We increased the FS of the qcow2 by extra 10G.

Currently rerunning the test on https://jenkins-continuous-infra.apps.ci.centos.org/job/fedora-f33-build-pipeline/40/

Would it be possible to make the disk size configurable in provision.fmf ?

Also, does each test within a yaml file run in a clean environment? Or do I need to put tests in separate yaml files to get a clean environment? I'm wondering if there is a way for me to split up the tests so that when a test is done it's artifacts aren't just sitting in the VM taking up disk space.

Would it be possible to make the disk size configurable in provision.fmf ?

Currently this is not possible, in case you need this feature, could you open an issue on https://pagure.io/standard-test-roles/ so we can discuss this there?

Also, does each test within a yaml file run in a clean environment? Or do I need to put tests in separate yaml files to get a clean environment? I'm wondering if there is a way for me to split up the tests so that when a test is done it's artifacts aren't just sitting in the VM taking up disk space.

The pipeline supports tests*.yml playbook, each playbook will run on fresh environment[ 1], so you could run each test case on separate playbook and each test case would run on fresh environment.

[1] https://docs.fedoraproject.org/en-US/ci/standard-test-interface/#_invocation

@tstellar can we close this issue?

Metadata Update from @bgoncalv:
- Issue status updated to: Closed (was: Open)

Metadata