Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.

Bug 1965026

Summary: [master] [Assisted-4.8][SNO][Dual Stack] DHCPv6 settings "ipv6.dhcp-duid=ll" missing from dual stack install
Product: OpenShift Container Platform Reporter: davis phillips <dphillip>
Component: assisted-installerAssignee: vemporop
assisted-installer sub component: assisted-service QA Contact: Yuri Obshansky <yobshans>
Status: CLOSED CURRENTRELEASE Docs Contact:
Severity: low    
Priority: high CC: aos-bugs
Version: 4.8Keywords: Triaged
Target Milestone: ---   
Target Release: 4.9.0   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard: AI-Team-Core KNI-EDGE-4.9
Fixed In Version: OCP-Metal-v1.0.21.3 Doc Type: No Doc Update
Doc Text:
Story Points: ---
Clone Of:
: 1966952 (view as bug list) Environment:
Last Closed: 2022-08-28 08:45:59 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:
Bug Depends On:    
Bug Blocks: 1966952    

Description davis phillips 2021-05-26 14:54:28 UTC
Description of problem:
Installation using CRs with dual stack fails to present persistent IPv6 DHCP requests. 

Version-Release number of selected component (if applicable):
4.8

How reproducible:
Every time


Steps to Reproduce:
1. Create dual stack deployment CR
Define CR for dual stack deployment:

apiVersion: extensions.hive.openshift.io/v1beta1
kind: AgentClusterInstall
metadata:
  name: dual-ocp-aci 
  namespace: assisted-installer
  annotations: {"agent-install.openshift.io/install-config-overrides": "{\"networking\": {\"networkType\": \"OVNKubernetes\", \"machineNetwork\": [{\"cidr\": \"10.19.32.192/26\"}, {\"cidr\": \"2620:52:0:1343::/64\"}], \"clusterNetwork\": [{\"cidr\": \"10.128.0.0/14\", \"hostPrefix\": 23}, {\"cidr\": \"fd01::/48\", \"hostPrefix\": 64}], \"serviceNetwork\": [\"172.30.0.0/16\", \"fd02::/112\"]}}"}
spec:
  clusterDeploymentRef:
    name: dual-ocp-sno
  imageSetRef:
    name: openshift-4.8 
  networking:
    clusterNetwork:
      - cidr: "10.128.0.0/14" 
        hostPrefix: 23
    serviceNetwork:
      - "172.30.0.0/16"
    machineNetwork:
      - cidr: "10.19.32.192/26"


2. Deploy cluster. Look for 

ipv6.dhcp-duid=ll

Here:
https://github.com/openshift/assisted-service/blob/7dac3fb956cc7c477cf79aa024ba4813ac85000e/internal/ignition/ignition.go#L809

AI should parse and return when an IPv6 address is found. 

https://github.com/openshift/assisted-service/blame/902a54d507dc4661d0ca2977114dc8500f52ee05/internal/network/utils.go#L22-L26


Actual results:
DHCPv6 sends IP address via DUID that includes timestamp and will change. 

Expected results:
DHCPv6 sends IP address via MAC address that will not change and can be reserved. 

Additional info:
Can provide kubeconfig for dual stack install if need be.

Comment 1 Ronnie Lazar 2021-05-26 15:33:44 UTC
Currently we do not support dual stack.
We will add support for it for 4.9

Comment 2 Ronnie Lazar 2021-05-26 15:35:53 UTC
Will be handled as part of https://issues.redhat.com/browse/MGMT-6670

Comment 4 vemporop 2021-06-03 11:22:20 UTC
@dphillip How does this problem affect installation and/or cluster behavior? Does cluster installation fail? The cluster isn't functional? I applied the fix and was able to finish installation (SNO, dual stack), but the cluster failed to finalize, with a few operators failed to become available.

Comment 5 davis phillips 2021-06-07 15:59:26 UTC
There is currently an MCO bug: https://bugzilla.redhat.com/show_bug.cgi?id=1966602

The MCO fails to come online because the kubelet configuration gets modified via: https://github.com/openshift/machine-config-operator/pull/2580

The problem is if the cluster is primarily IPv6 and SNO without the fix here. The IP address assigned via DHCPv6 is unreliable and changes frequently. 

Did you test with then api DNS endpoint being IPv6?

Comment 6 davis phillips 2021-06-10 19:34:24 UTC
Closed. Verified