When triggerring a pipeline automatically (with a trigger token, probably doesn’t matter), outdated/old interruptible
jobs are cancelled and marked as “failed”.
Every time this happens I get a notification for a failure and then a notification that my pipeline was fixed (after a successful run). This is quite a problem as I’m building a static site which gets built automatically (via trigger token) every time the data changes.
Here’s my gitlab-ci.yml
:
build:
stage: build
image: docker:19.03.11
services:
- docker:19.03.11-dind
variables:
DOCKER_HOST: tcp://docker:2375
DOCKER_DRIVER: overlay2
IMAGE_TAG: $CI_REGISTRY_IMAGE
before_script:
- docker login -u $CI_REGISTRY_USER -p $CI_REGISTRY_PASSWORD $CI_REGISTRY
script:
- docker pull $IMAGE_TAG || true
- docker build --tag $IMAGE_TAG .
- docker push $IMAGE_TAG
interruptible: true
deploy:
stage: deploy
image: appropriate/curl
variables:
WEBHOOK: $DEPLOY_WEBHOOK # workaround for escaping `$`
script:
- curl $WEBHOOK -H "" -d ""
Thanks in advance!