Workflow: openshift-e2e-gcp-ccm

The Openshift E2E GCP workflow using CCM as a primary mean to initialize nodes and create external LoadBalancers. Executes the common end-to-end test suite on GCP to test updated cluster configuration.

Pre Steps

Name Description
ipi-conf The IPI configure step generates the generic portions of the install-config.yaml file based on optional input files.
ipi-conf-gcp This step configures an install-config.yaml file to depoloy to the GCP CI project. An existing install-config.yaml file should already exist in the shared directory. The file is modified to configure the cluster to run in the GCP CI project. The resulting file remains in the shared directory for future consumption.
ipi-install The IPI install step chain contains all the individual steps necessary to install an OpenShift cluster.
ccm-conf-apply-feature-gate Apply custom featureGate configuration to cluster to enable CCM functionality
storage-conf-csi-gcp-pd The csi-gcp-pd step creates ${SHARED_DIR}/${TEST_CSI_DRIVER_MANIFEST} for GCP PD CSI driver, which is consumed by openshift-e2e-test step to run CSI driver certification suite.
storage-conf-wait-for-csi-migration storage-conf-csi-migration-enable steps enables alpha feature gates for CSI migration of all in-tree volume plugins we have in OCP and waits for the chanegs to be reflected in kube-controller-manager and nodes. It expects that there is no volume present on the system that would be affected by the migration!

Test Steps

Name Description
openshift-e2e-test The Openshift E2E step executes the common end-to-end test suite.

Post Steps

Name Description
gather-core-dump The network step runs the must gather network tool to gather core dumps.
ccm-must-gather The must-gather step runs the must-gather tool to gather ccm artifacts.
ipi-gcp-post The IPI cleanup step contains all steps that deprovision an OpenShift cluster on GCP, provisioned by the `ipi-gcp-pre` chain.


Image Exposed As Override[?] Required By Steps
release:latest OPENSHIFT_INSTALL_RELEASE_IMAGE_OVERRIDE no ipi-install-install
RELEASE_IMAGE_LATEST no ipi-install-install


In addition to the default environment, the following variables are consumed through this workflow

Variable Name Variable Content Consumed By Steps
COMPUTE_NODE_TYPE The instance type to use for compute nodes (e.g. GCP We use a 4 core worker to match the median configuration of the fleet. (default: n1-standard-4) ipi-conf-gcp
CONTROL_PLANE_REPLICAS The number of control plane nodes to create for the cluster. This variable should only be supplied in consultation with the etcd team. The etcd operator has to explicitly support the specified number of control plane nodes and values not documented below will result in unhealthy and/or unsupportable clusters. Valid values include: * "" (default) - 3 control plane nodes. The standard for stand-alone clusters. * "1" - Supportable for single-node OpenShift (SNO). * "5" - Supportable only for select customers. (default: 3) ipi-conf-gcp
FIPS_ENABLED (default: false) ipi-conf
KCMO_IMAGE_OVERRIDE ccm-conf-apply-feature-gate
MCO_IMAGE_OVERRIDE ccm-conf-apply-feature-gate
OPENSHIFT_INSTALL_EXPERIMENTAL_DUAL_STACK Using experimental Azure dual-stack support (default: false) ipi-install-install
OPENSHIFT_INSTALL_PRESERVE_BOOTSTRAP Keep bootstrap resources after install. Warning: this should only be used for debugging purposes, and poses a risk to cluster stability. ipi-install-install
RT_ENABLED This flag forces the installation of real-time kernel to worker nodes. (default: false) ipi-conf-gcp
SIZE_VARIANT The size of the cluster in one of our supported t-shirt values that is standard across all CI environments. The sizes are: * "" (default) - 4 vCPU, 16GB control plane nodes, default workers * "compact" - 8 vCPU, 32GB control plane nodes, no workers * "large" - 16 vCPU, 64GB+ control plane nodes, default workers, suitable for clusters up to 250 nodes * "xlarge" - 32 vCPU, 128GB+ control plane nodes, default workers, suitable for clusters up to 1000 nodes These sizes are roughly consistent across all cloud providers, but we may not be able to instantiate some sizes in some regions or accounts due to quota issues. ipi-conf-gcp
TEST_ARGS Additional arguments to be passed to 'openshift-test' openshift-e2e-test
TEST_CSI_DRIVER_MANIFEST The CSI driver manifest to create. Used by the `openshift-tests` as a manifest for openshift/csi tests. The file must be in ${SHARED_DIR}. storage-conf-csi-gcp-pd openshift-e2e-test
TEST_INSTALL_CSI_DRIVERS The CSI driver(s) to use. Used by the `openshift-tests` program, see its documentation for details. openshift-e2e-test
TEST_REQUIRES_SSH If the test suite requires SSH access, setting this variable will install a bastion and configure the Kube e2e test infrastructure to use the bastion. Otherwise, tests that require SSH will fail because OpenShift does not expose SSH to nodes by default. openshift-e2e-test
TEST_SKIPS Regular expression (POSIX basic regular expression) of tests to skip. It is suggested to test the regex to make sure that it matches with the available tests. Tests can be listed by using 'openshift-tests run --dry-run (...)'. Sometimes, the tests that are printed in Prow won't exactly match the list returned by openshift-tests. openshift-e2e-test
TEST_SUITE The test suite to run. Use 'openshift-test run --help' to list available suites. (default: openshift/conformance/parallel) openshift-e2e-test
TEST_TYPE The type of test to perform. * 'suite' - Run the test suite defined by TEST_SUITE. By default this executes the OpenShift parallel conformance suite. All product components must run the default suite on pull requests prior to merge. * 'suite-conformance' - Runs the test suite defined by TEST_SUITE, then runs the product conformance suite to verify the cluster is still functional. This is typically used for testing disruption from one suite, then confirming that the disruption did not degrade the product. * 'upgrade' - Perform an upgrade to the image defined by OPENSHIFT_UPGRADE_RELEASE_IMAGE_OVERRIDE. The TEST_UPGRADE_OPTIONS flag may be optionally set. All product components must run the default upgrade test (verify an upgrade completes) on pull requests prior to merge. * 'upgrade-conformance' - Performs the 'upgrade' test and then executes the full conformance suite after upgrade completes. Does not honor TEST_SUITE but will respect TEST_UPGRADE_OPTIONS. All product releases must pass the conformance suite after an upgrade completes or have explicit logic in their test to tolerate behavior after upgrade. * 'upgrade-paused' - Perform an upgrade to the images defined by OPENSHIFT_UPGRADE_RELEASE_IMAGE_OVERRIDE with paused worker pool. (default: suite) openshift-e2e-test
TEST_UPGRADE_OPTIONS Options controlling how an upgrade is performed. See `openshift-test run-upgrade --help` for more details. openshift-e2e-test
TEST_UPGRADE_SUITE The test suite to run in upgrade tests. Use 'openshift-test run --help' to list available suites. (default: all) openshift-e2e-test
TMPDIR A pathname of a directory made available for programs that need a place to create temporary files. (default: /tmp) gather-gcp-console

Step Graph

Webreg Workflow "openshift-e2e-gcp-ccm" cluster_1 Pre cluster_0 ipi-install cluster_2 Test cluster_7 Post cluster_3 gather-core-dump cluster_6 ipi-gcp-post cluster_5 ipi-deprovision cluster_4 gather 0 ipi-conf 1 ipi-conf-gcp 0->1 2 ipi-install-rbac 1->2 3 openshift-cluster-bot-rbac 2->3 4 ipi-install-install 3->4 5 ipi-install-times-collection 4->5 6 ccm-conf-apply-feature-gate 5->6 7 storage-conf-csi-gcp-pd 6->7 8 storage-conf-wait-for-csi-migration 7->8 9 openshift-e2e-test 8->9 10 gather-core-dump 9->10 11 ccm-must-gather 10->11 12 gather-gcp-console 11->12 13 gather-must-gather 12->13 14 gather-extra 13->14 15 gather-audit-logs 14->15 16 ipi-deprovision-deprovision 15->16

GitHub Link:




Source code for this page located on GitHub