In case `wget` is not installed, downloading the Helm release will fail.
The `install-helm.sh` script won't return a fatal error in that case,
and CI jobs continue running in an environment that is not ready.
By adding a check that exist the script with a failure, the CI will now
correctly report a problem when Helm can not be downloaded.
See-also: #3430
Signed-off-by: Niels de Vos <ndevos@redhat.com>
When running on AWE EC2 virtual-machines, we'll use Podman instead of
installing a VM. The "none" driver might work as well, but it requires
additional dependencies to be installed, which may change over time with
new minikube or Kubernetes releases. Hopefully the Podman driver is less
affected with changes in dependencies.
Depends-on: #3419Closes: #3415
Signed-off-by: Niels de Vos <ndevos@redhat.com>
as PSP is deprecated in kubernetes 1.21
and will be removed in kubernetes 1.25
removing the existing PSP related templates
from the repo and updated the required documents.
fixes#1988
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
All the CSI spec imports are having snakecase
we cannot fix it, instead of disabling all imports
skipping the nosnakecase static check.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
At present, the check is performed to validate the version of kube
is v1.17 and this commit remove the same.
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
The Ceph v17.2.2 container-image fails to start Ceph Mgr. This causes
issues while the e2e test suite is running. It is better to check if
Ceph Mgr is available, before continuing with the rest of the CI job.
Updates: #3259
Signed-off-by: Niels de Vos <ndevos@redhat.com>
PodSecurity featuregate is beta in kubernetes
1.23 and its causing problem for the existing
tests. This PR disables the PodSecurity featuregate
for now and will be enabled later.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
Still seeing the issue of the commitlint
as below
fatal: unsafe repository
('/go/src/github.com/ceph/ceph-csi'
is owned by someone else)
To add an exception for this directory,
call:
git config --global --add safe.directory \
/go/src/github.com/ceph/ceph-csi
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
`bash -E` causes inheritance of the ERR trap into shell functions,
command substitutions, and commands executed in a subshell environment.
Because the `kubectl_retry` function depends on detection an error of a
subshell, the ERR trap is not needed to be executed. The trap contains
extra logging, and exits the script in the `rook.sh` case. The aborting
of the script is not wanted when a retry is expected to be done.
While checking for known failures, the `grep` command may exit with 1,
if there are no matches. That means, the `ret` variable will be set to
0, but there will also be an error exit status. This causes `bash -E` to
abort the function, and call the ERR trap.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
Depending on the Kubernetes version, the following warning is reported
regulary:
> Warning: policy/v1beta1 PodSecurityPolicy is deprecated in v1.21+,
> unavailable in v1.25+
The warning is written to stderr, so skipping AlreadyExists or NotFound
is not sufficient to trigger a retry. Ignoring '^Warning:' in the stderr
output should prevent unneeded failures while deploying Rook or other
components.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
Rook deployments fail quite regulary in the CI environment now. It is
not clear what the cause is, hopefully a little better logging will
guide us to the issue.
Now executing `kubectl` in a sub-shell, ensuring that the redirection of
the command lands in the right files.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
This commit change the image registry URL for sidecars in the
deployment from `k8s.gcr.io` to `registry.k8s.io` as
the migration is happening from former to the latter. This commit
also correct the e2e readme for the change.
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
Considering snapshot controllers have been moved to GA since
kube version 1.20, we no longer need to have a mention of beta
version of the same in our deployment.
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
The sidecar images in minikube deployment will be fetched from
build.env and used/validated accordingly.
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
The CentOS 8 repository for Apache Arrow has been removed. This causes
container-image builds fail with the following error:
Errors during downloading metadata for repository 'apache-arrow-centos':
- Status code: 404 for https://apache.jfrog.io/artifactory/arrow/centos/8/x86_64/repodata/repomd.xml (IP: 54.190.66.70)
Error: Failed to download metadata for repo 'apache-arrow-centos': Cannot download repomd.xml: Cannot download repodata/repomd.xml: All mirrors were tried
The Ceph base image has `arrow/centos/8` configured, maybe Apache Arrow
offers a CentOS Stream 8 repository now? Once the Ceph container-image
has been updated, the repository can be enabled again.
Ceph-CSI does not depend on Apache Arrow, so there is no functional
change by disabling the repository.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
in Rook 1.8 templates the failureDomain is
set to host for ec pools, as we are using
single node minikube cluster, setting the
failureDomain for osd.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
Rook creates a detault pool with name
device_health_metrics for
device-health-metrics CephBlockPool CR.
device-health-metrics is added to cluster-test.yaml
in Rook 1.8
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
kubectl log with labels will log only
last 10 lines by default adding tail=-1
to log complete output.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
In Rook v1.8+ the path for the deployment articafts
are changed from `"cluster/examples/kubernetes/ceph`
to `deploy/examples`.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
currently, rook.sh code is not aligned properly
unside the functions, this commit does
the code alignment.
PS: this is done by vscode for me :)
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
we dont need to specify the image version
separately in minikube script, use the image
version defined in the build.env
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
The ReadWriteOncePod feature gate need to be enabled only when we
are operating on kube 1.22 or above cluster. This commit adds the
logic to parse the kubernetes cluster at time of minikube deployment
and if it is above v1.22, enable the RWOP feature gate
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
The golangci-lint install script at goreleaser.com is deprecated. Docs
now advise to install from a github link:
goreleaser/godownloader#207
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
Instead of installing the amd64 on all the
platforms, install architecture specific go
version in test dockerfile.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
Instead of installing the amd64 on all the
platforms, install architecture specific go
version for devel dockerfile
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
currently getting find command not found
and xargs command not found when we run
the dockerfile.test. installing findutils
to fix it.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
The `expandCSIVolumes` feature gate is beta since kubernetes 1.16
version and we no longer wanted to explictly enable it in the
deployment.
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
as there are lot of spell check failures
on the vendor directory, skipping it.
skipping retest action folder as
PullRequests word is not getting ignored
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
