-
Bug
-
Resolution: Done-Errata
-
Normal
-
4.10.z
-
None
-
Moderate
-
No
-
1
-
Sprint 242, Sprint 243
-
2
-
Rejected
-
False
-
This is a clone of issue OCPBUGS-4359. The following is the description of the original issue:
—
Description of problem:
when fs full, update-dns-resolver fails to build a proper /etc/hosts, resulting in /etc/hosts only containing the openshift-generated-node-resolver lines, missing the localhost lines. This causes issues on pods having hostNetwork: true like openstack-cinder-csi-driver-controller
Version-Release number of selected component (if applicable):
OpenShift 4.10.39
How reproducible:
See: https://github.com/openshift/cluster-dns-operator/blob/a5ea3fcb7be49a12115bd6648403df3d65661542/assets/node-resolver/update-node-resolver.sh
Steps to Reproduce:
1. make sure the file system is full when running the cp at line 13 2. 3.
Actual results:
/etc/hosts is missing the localhost lines
Expected results:
/etc/hosts should contain the localhost lines
Additional info:
- causes
-
OCPBUGS-18654 kube-apiserver was flooded by localhost DNS lookup connection error
- Closed
- clones
-
OCPBUGS-4359 cluster-dns-operator corrupts /etc/hosts when fs full
- Closed
- is blocked by
-
OCPBUGS-4359 cluster-dns-operator corrupts /etc/hosts when fs full
- Closed
- links to
-
RHBA-2023:5677 OpenShift Container Platform 4.12.z bug fix update
(1 links to)