-
Bug
-
Resolution: Done
-
Undefined
-
4.12.z, 4.14.z, 4.15.z, 4.17.z, 4.16.z, 4.18.z, 4.19.z
-
None
-
Quality / Stability / Reliability
-
False
-
-
5
-
Moderate
-
None
-
None
-
None
-
None
-
OSDOCS Sprint 272, OSDOCS Sprint 273
-
2
-
None
-
None
-
None
-
None
-
None
-
None
-
None
Description of problem:
Somewhere we recommend the DNS rotate option to try to improve DNS performance.
This option is counter to the DNS design of OpenShift CoreDNS and breaks api-int lookups inside the cluster.
dns-resolver: config: options: - rotate
We should add a warning to the NMstate DNS docs to prevent users from using the rotate option.
Maybe somewhere around modules/ipi-install-configuring-host-network-interfaces-in-the-install-config.yaml-file.adoc
Version-Release number of selected component (if applicable):
4.16.38
How reproducible:
Once
Steps to Reproduce:
1. Install a cluster with the above NMstate with 'rotate' option 2. Check multus pod api-int DNS lookups 3.
Actual results:
DNS lookups fail, pods stuck in ContainerCreating, multis CNI adds fail, ovnkube-node pods CrashLoopBackOff
E0609 18:40:35.256778 86824 reflector.go:147] k8s.io/client-go/informers/factory.go:159: Failed to watch *v1.Pod: failed to list *v1.Pod: Get "https://api-int.cluster.example.com:6443/api/v1/pods?fieldSelector=spec.nodeName%3Dworker-01.cluster.example.com&resourceVersion=36232299": dial tcp: lookup api-int.cluster.example.com: no such host
Expected results:
api-int.<cluster_name>.<base_domain> lookups succeed from inside the cluster.
Additional info:
Possible old KB.
How to tune the DNS client resolver library through /etc/resolv.conf
- documents
-
OCPBUGS-57343 openshift-kube-apiserver_installer-20 pods are stuck in ContainerCreating due to a recurring Multus CNI error: "error waiting for pod: pod not found".
-
- Closed
-
- links to