Move the mini-e2e job into a template-job and generate two jobs out of
it: mini-e2e/k8s-1.17.8 and mini-e2e/k8s-1.18.5
By passing the k8s_version as variable to the job-template, and placing
it in the parameters for the mini-e2e.groovy script, all hard-coded
occurences of the Kubernetes version can be replaced by the
{k8s_version} placeholder.
See-also: https://jenkins-job-builder.readthedocs.io/en/latest/definition.html#job-template
Signed-off-by: Niels de Vos <ndevos@redhat.com>
Commit f5cba3aaa8 added the mini-e2e job, but still referred to the
temporary location that was used for testing the job. As everything is
available in the ceph-csi:ci/centos repository:branch, there is no need
to refer to the temporary location.
Reported-by: Yug <yuggupta27@gmail.com>
Signed-off-by: Niels de Vos <ndevos@redhat.com>
While debugging issues with the job itself, a sleep has been very
useful. PRs that have been rebased on the master branch contain all the
deployment fixes that are needed for the job to pass. There is no need
anymore to run into the long sleep when the job fails.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
Restarting the CI job can now be done with the following comments in a
PR:
- /test all
- /retest all
- /test ci/centos/mini-e2e
- /retest ci/centos/mini-e2e
- /test ci/centos/mini-e2e/k8s-1.18
- /retest ci/centos/mini-e2e/k8s-1.18
This is added as preparation for enabling running the mini-e2e job with
different Kubernetes versions.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
The CentOS CI jobs that use single-node-k8s.sh will run minikube with
additional disks. That makes it possible for Rook 1.3 and newer to
deploy the Ceph cluster.
The current Travis CI jobs run minikube on the Travis provided
environment, without any extra disks. This prevents updating Rook for
all CI jobs at the moment.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
The new mini-e2e jobs does the following:
- reserve a bare-metal machine
- checkout the git repository with the PR
- build used artifacts (container image and e2e.test executable)
- deploy k8s and Rook in a minikube VM
- run the e2e tests
With-contributions-from: Yug <yuggupta27@gmail.com>
Signed-off-by: Niels de Vos <ndevos@redhat.com>
Currently CentOS-7 machines were requested. CentOS-8 has been out for a
while now, and is stable for running manual jobs. There is nothing
preventing us from using CentOS-8 bare-metal machines for testing.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
On the new OCP4 cluster, deleting the batch job results in deletion of
the pod as well. Deleting the pod when it does not exist anymore results
in a failure. Ignoring the potential failure makes the job pass again.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
Currently the number of Pods is growing in the OpenShift Console. The
pods are all Completed, so not running anymore, but there is also no
need to keep them around.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
With the jjb/session label on the Pod, tracking the job works again.
Therefor this test can be re-enabled.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
By using a template, it becomes possible to identify the Pod that has
been started by the Batch Job.
This prevents the script from getting the logs from an incorrect (old)
container.
Fixes: #1111
Signed-off-by: Niels de Vos <ndevos@redhat.com>
The jjb-validate job is not correct, it does not always return the
results of the PR.
Prevent triggering the job automatically, and only have it run when
someone leaves the '/test ci/centos/jjb-validate' comment.
Updates: #1111
Signed-off-by: Niels de Vos <ndevos@redhat.com>
Using the 'git' function in groovy does not allow checking out a
non-branch, so use the 'checkout' function instead.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
The 'scm' section also needs to be a list. Jenkins Jobs Builder does not
detect this while validating, but it cause an error while updating the
jobs in Jenkins.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
In case there is no 'origin', jenkins-jobs fails with an error like:
jenkins_jobs.errors.JenkinsJobsException: Must specify a url for git remote "origin"
Signed-off-by: Niels de Vos <ndevos@redhat.com>
The directory that gets constructed can be like
`/opt/build/deploy/..//jobs` and this causes jenkins-jobs to fail.
Enternig the WORKDIR and passing the local ./jobs directory to
jenkins-jobs makes things clearer.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
There is a new .commitlintrc.yml file in the root of the repository.
This causes jenkins-job-builder to complains that the file is not
confirm the Jenkins Job Builder format/specification.
jenkins_jobs.errors.JenkinsJobsException: The topmost collection in file '/opt/build/.commitlintrc.yml' must be a list, not a <class 'collections.OrderedDict'>
Move all Jenkins Job Builder .yaml files to jobs/ so that all the
Jenkins job configurations are kept together without other files.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
The `git` command does not allow checking out a non-branch (like a PR),
so it is needed to use `checkout` instead.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
When checking out a PR for a non-master branch, `git clone` should
download the last commit of the branch. Adding a `--base=..` option to
pass the cloning of a selected branch, instead of `master`.
We also want to fetch all commits in the PR, so they can get tested with
`commitlint`. Only fetch --depth=1 the initial clone, but fetch
everything in the PR itself.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
This job runs 'make' on the ci/centos branch (or Pull Request) so that
all MarkDown, shell-scripts and yaml files are validated.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
This merges several files from the master branch into ci/centos so that
the scripts, MarkDown and yaml files can be checked.
Run 'make' or 'make test' to validate the contents of this branch.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
./README.md:51: MD012 Multiple consecutive blank lines
./README.md:10: MD025 Multiple top level headers in the same document
Signed-off-by: Niels de Vos <ndevos@redhat.com>
20:11 error wrong indentation: expected 12 but found 10 (indentation)
22:11 error wrong indentation: expected 12 but found 10 (indentation)
Signed-off-by: Niels de Vos <ndevos@redhat.com>
16:7 error wrong indentation: expected 8 but found 6 (indentation)
19:9 error wrong indentation: expected 10 but found 8 (indentation)
Signed-off-by: Niels de Vos <ndevos@redhat.com>
./deploy/jjb-deploy.yaml
16:7 error wrong indentation: expected 8 but found 6 (indentation)
19:9 error wrong indentation: expected 10 but found 8 (indentation)
26:9 error wrong indentation: expected 10 but found 8 (indentation)
30:7 error wrong indentation: expected 8 but found 6 (indentation)
Signed-off-by: Niels de Vos <ndevos@redhat.com>
In ./deploy/jjb.sh line 11:
oc get pod/${1} --no-headers -o=jsonpath='{.status.phase}'
^--^ SC2086: Double quote to prevent globbing and word splitting.
In ./deploy/jjb.sh line 29:
cd $(dirname ${0})
^----------------^ SC2164: Use 'cd ... || exit' or 'cd ... || return' in case cd fails.
^-------------^ SC2046: Quote this to prevent word splitting.
^--^ SC2086: Double quote to prevent globbing and word splitting.
In ./deploy/jjb.sh line 31:
oc create -f jjb-${CMD}.yaml
^----^ SC2086: Double quote to prevent globbing and word splitting.
In ./deploy/jjb.sh line 36:
jjb_pod=$(oc get pods --no-headers -l job-name=jjb-${CMD} -o=jsonpath='{.items[0].metadata.name}')
^----^ SC2086: Double quote to prevent globbing and word splitting.
In ./deploy/jjb.sh line 37:
[ ${?} = 0 ] && [ -n "${jjb_pod}" ] && break
^--^ SC2181: Check exit code directly with e.g. 'if mycmd;', not indirectly with $?.
In ./deploy/jjb.sh line 46:
[ ${?} = 0 ] && ( [ "${status}" = "Succeeded" ] || [ "${status}" = "Failed" ] ) && break
^--^ SC2181: Check exit code directly with e.g. 'if mycmd;', not indirectly with $?.
^-- SC2235: Use { ..; } instead of (..) to avoid subshell overhead.
In ./deploy/jjb.sh line 54:
oc delete --wait -f jjb-${CMD}.yaml
^----^ SC2086: Double quote to prevent globbing and word splitting.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
In ./prepare.sh line 21:
if [ $? -ne 0 ]
^-- SC2181: Check exit code directly with e.g. 'if mycmd;', not indirectly with $?.
In ./prepare.sh line 28:
eval set -- ${opts}
^-----^ SC2086: Double quote to prevent globbing and word splitting.
In ./prepare.sh line 55:
echo ${ref}
^----^ SC2086: Double quote to prevent globbing and word splitting.
In ./prepare.sh line 69:
git clone --depth=1 ${gitrepo} ${workdir}
^--------^ SC2086: Double quote to prevent globbing and word splitting.
^--------^ SC2086: Double quote to prevent globbing and word splitting.
In ./prepare.sh line 70:
cd ${workdir}
^-----------^ SC2164: Use 'cd ... || exit' or 'cd ... || return' in case cd fails.
^--------^ SC2086: Double quote to prevent globbing and word splitting.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
To prevent the failure of job due to unavailability
of a machine immediately, retry mechanism is used.
If unable to reserve a machine, it will retry every
5 mins for 30 times to avoid job failure.
Signed-off-by: Yug Gupta <ygupta@redhat.com>
These additional settings are required to trigger the job to run when a
GitHub PR gets created or updated. It is also possible to manually run
the job by leaving one of the following comments:
/test ci/centos/containerized-tests
/retest ci/centos/containerized-tests
/test all
/retest all
The 'all' has been added so that future jobs can catch this comment as
well and act on it.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
