NVIDIA / gpu-operator / 41014572133
37%
main: 37%

LAST BUILD BRANCH: pull-request/3089
DEFAULT BRANCH: main
Ran
Jobs 1
Files 127
Run time 1min
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 37.155%. First build
41014572133

push

github

abrarshivani 
ci: use the newest operand image available for the main build

The daily main build took each operand's newest commit and failed when
that commit had no image, so one operand whose CI broke stopped the
whole build. On Oct 29 and 30 the container toolkit's newest commit had
no image because its centos7 package job failed, and on Oct 30 the
kubevirt GPU device plugin's newest commits had none because they
failed the notices check.

get-latest-images.sh now reads the last 30 commits of each operand's
branch and uses the newest one that has an image. It no longer waits
for the newest commit's image: a commit merged minutes before the run
falls back to the one before it. When it falls back, it prints a
warning naming the image, the commit it was built from and that
commit's date, with links to the commit and to the newer commits on the
branch, and it still fails when none of the 30 commits has an image.
Only a 404 from the registry counts as a missing image; any other
registry error is retried twice and then fails the lookup rather than
falling back.

Images taken from a commit older than 24 hours because newer commits
have no image are written to the file given with
--outdated-images-file. publish-main-chart.yaml posts them to Slack
after a scheduled build is recorded, so a stale operand is noticed
without blocking the build.

Signed-off-by: Abrar Shivani <ashivani@nvidia.com>

6015 of 16189 relevant lines covered (37.15%)

0.37 hits per line

Jobs
ID Job ID Ran Files Coverage
1 41014572133.1 127
37.15
GitHub Action Run
Source Files on build 41014572133