Bug 2102739
| Summary: | Baremetal IPI Installation with ipv6 fails on because of missing resolv.conf on masters | ||
|---|---|---|---|
| Product: | OpenShift Container Platform | Reporter: | Karim Boumedhel <kboumedh> |
| Component: | Installer | Assignee: | Ben Nemec <bnemec> |
| Installer sub component: | OpenShift on Bare Metal IPI | QA Contact: | Amit Ugol <augol> |
| Status: | CLOSED DUPLICATE | Docs Contact: | |
| Severity: | high | ||
| Priority: | urgent | CC: | bnemec, dmellado, mavazque, shardy |
| Version: | 4.11 | ||
| Target Milestone: | --- | ||
| Target Release: | --- | ||
| Hardware: | All | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | If docs needed, set a value | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2022-07-11 21:34:00 UTC | Type: | Bug |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
|
Description
Karim Boumedhel
2022-06-30 13:53:15 UTC
i feel it has a relation with the file /etc/NetworkManager/conf.d/99-kni.conf, with the following content [main] rc-manager=unmanaged [connection] ipv6.dhcp-duid=ll ipv6.dhcp-iaid=mac Could you please provide some logs? I assume must-gather is not possible, maybe journalctl, container logs, anything? How did you come to the conclusion about 99-kni.conf? the same file seems to work fine with previous releases, so I rather think it's a change of behaviour in NetworkManager (1.36 in rhcos4.11 vs 1.30 in rhcos4.10) i can provide a live environment with the issue when needed, just ping me > i feel it has a relation with the file /etc/NetworkManager/conf.d/99-kni.conf, with the following content You may be right, but as already mentioned we'll need logs or some live-debug to figure out what's wrong. The `rc-manager=unmanaged` is because we have a NM dispatcher script that discovers the DNS servers via DHCP, then prepends localhost for the local coredns service (which resolves cluster-internal names) https://github.com/openshift/machine-config-operator/blob/master/templates/common/on-prem/files/NetworkManager-resolv-prepender.yaml To see the logs for that script check the NetworkManager-dispatcher.service journal on the masters, if you can attach the logs here we may be able to spot some clues, thanks! using the following machineconfig (to set rc-manager=file) allows the install to work just fine
```
apiVersion: machineconfiguration.openshift.io/v1
kind: MachineConfig
metadata:
name: empty-99-kni
labels:
machineconfiguration.openshift.io/role: master
spec:
config:
ignition:
version: 2.2.0
storage:
files:
- contents:
source: data:text/plain;charset=utf-8;base64,W21haW5dCnJjLW1hbmFnZXI9ZmlsZQpbY29ubmVjdGlvbl0KaXB2Ni5kaGNwLWR1aWQ9bGwKaXB2Ni5kaGNwLWlhaWQ9bWFjCg==
verification: {}
filesystem: root
mode: 420
path: /etc/NetworkManager/conf.d/99-kni.conf
```
This seems to be a duplicate of https://bugzilla.redhat.com/show_bug.cgi?id=2105069. Ben, can you please confirm and close it as such if so? I've assigned it to you as it seems to be within the metal networking wheelhouse. Please flip it back if incorrect. Yeah, sorry, I meant to mark this as duplicate when I first got pinged. *** This bug has been marked as a duplicate of bug 2105069 *** |