deads2k

{
  "apiVersion": "testing.openshift.io/v1",
  "kind": "SSAWithSet",
  "metadata": {
    "creationTimestamp": "2024-04-24T20:58:42Z",
    "generation": 1,
    "managedFields": [
      {
        "apiVersion": "testing.openshift.io/v1",
        "fieldsType": "FieldsV1",

deads2k

        {
            "GroupVersionResource": {
                "Group": "config.openshift.io",
                "Version": "v1",
                "Resource": "clusteroperators"
            },
            "RequestCounts": {
                "RequestStartedCount": 0,
                "RequestFinishedCount": 25256,
                "ClientFailedRequestCount": 1716,

deads2k

how do we ensure only a resource is managed by at most one operator install.

naivest approach:  **This one is subpar**
1. in-process controller determines list of all resources to be created
2. check each resource in the controller to see if it is owned by something else 
   (probably a list across all namespaces of something (extensions?) )
3. if there is another owner, the contorller refuses to create the resource

When I install an app I get a namespaced extension resource
1. extension resource exists in the namespace I installed the app in

deads2k

apiVersion
kind: NeverCreated
metadata
  annotations:
    "include.release.openshift.io/ibm-cloud-managed": false
spec:
  featureGateTests:
  - featureGateName: Example
    tests:
    - "[sig-arch][OCPFeatureGate:Example] should only run FeatureGated test when enabled"

deads2k

FeatureGate	Default on Hypershift	Default on SelfManagedHA	Default on SingleNode	LatencySensitive on Hypershift	LatencySensitive on SelfManagedHA	LatencySensitive on SingleNode	TechPreviewNoUpgrade on Hypershift	TechPreviewNoUpgrade on SelfManagedHA	TechPreviewNoUpgrade on SingleNode
BareMetalLoadBalancer	Adding as Enabled	Adding as Enabled	Adding as Enabled	Not Available	Not Available	Not Available	Adding as Enabled	Adding as Enabled	Adding as Enabled
KMSv1

deads2k

k8s.io/apiserver
  rapid-reset, defaulted to off.  In repos vendoring, I needed to enable/disable

k8s.io/client-go
  aggregated-discovery - had a bug and needed  way for every vendoring binary we cannot recompile to disable. We added env vars.
  streaming-list - desire to have different default in kube-controller-manager versus kubelet

kube-controller-manager 
  client-go/streaming-list - disabled by default in library, kcm wants it enabled by default and controlled via flag.
  

deads2k

func init() {
  registerOpenshiftFeatures()
	runtime.Must(utilfeature.DefaultMutableFeatureGate.Add(defaultKubernetesFeatureGates))
}

deads2k

Logs for -n e2e-pod-network-disruption-test-k92nt pod/pod-network-to-service-disruption-poller-65d867cdbd-drk7j
Initializing to watch clusterIP 172.30.19.181:80
Initializing to watch clusterIP 172.30.136.5:80
Watching configmaps...
Adding and starting: http://172.30.136.5:80 on node/ci-op-89h3f2lm-9825d-5bzsl-worker-c-slpbx
Successfully started: http://172.30.136.5:80 on node/ci-op-89h3f2lm-9825d-5bzsl-worker-c-slpbx
Interrupted, terminating
Waiting for watchers to close...
{"level":"Info","locator":"backend-disruption-name/pod-to-service-reused-connections connection/reused disruption/pod-to-service-to-service-from-node-ci-op-89h3f2lm-9825d-5bzsl-worker-c-slpbx-to-clusterIP-172.30.136.5","message":"reason/DisruptionEnded backend-disruption-name/pod-to-service-reused-connections connection/reused disruption/pod-to-service-to-service-from-node-ci-op-89h3f2lm-9825d-5bzsl-worker-c-slpbx-to-clusterIP-172.30.136.5 started responding to GET requests over reused connections","tempSource":"Disruption","tempStructured

deads2k

bash: alias: `cd-github.com/deads2k/structured-merge-diff': invalid alias name
deads@fedora:~/workspaces/aws-ebs-csi-driver-operator/src/github.com/openshift/aws-ebs-csi-driver-operator$ cd ~/workspaces/jupierce/intervals-viewer/
deads@fedora:~/workspaces/jupierce/intervals-viewer$ source venv/bin/activate
(venv) deads@fedora:~/workspaces/jupierce/intervals-viewer$ pip install -r requirements.txt
Collecting arcade==2.6.17 (from -r requirements.txt (line 1))
  Using cached arcade-2.6.17-py3-none-any.whl (39.4 MB)
Collecting pandas==2.2.0 (from -r requirements.txt (line 2))
  Obtaining dependency information for pandas==2.2.0 from https://files.pythonhosted.org/packages/b2/a5/7f14d11f5bb3ca5681f6827616ccfbb03ec9504322674e4f962a5e9e404b/pandas-2.2.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata
  Using cached pandas-2.2.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (19 kB)
Collecting requests==2.31.0 (from -r requirements.txt (line 3))

deads2k

from https://gcsweb-ci.apps.ci.l2s4.p1.openshiftapps.com/gcs/test-platform-results/logs/periodic-ci-openshift-release-master-ci-4.16-e2e-aws-sdn-serial/1750498185056358400/artifacts/e2e-aws-sdn-serial/gather-audit-logs/artifacts/

deads@fedora:~/workspaces/cluster-debug-tools/src/github.com/openshift/cluster-debug-tools$ ./kubectl-dev_tool audit -f '/home/deads/Downloads/audit-logs(4)/9e67a938c529c2e045643d88bb48f105d0d45369e4e535d8e3b1163a47e22dca/audit_logs/kube-apiserver'   -otop --by=verb --user=system:serviceaccount:openshift-console-operator:console-operator
had 8600 line read failures
count: 36796, first: 2024-01-25T08:02:24-05:00, last: 2024-01-25T10:12:01-05:00, duration: 2h9m36.97217s

Top 10 "CREATE" (of 41 total hits):
     23x [  10.92813ms] [409-22] /apis/console.openshift.io/v1/consolenotifications                                      [system:serviceaccount:openshift-console-operator:console-operator]
      5x [    6.9752ms] [201-4]  /api/v1/namespaces/openshift-console/configmaps