Bug 1967632 - NetworkManager gives up on dhcp4 when it gets dhcp6 lease
Summary: NetworkManager gives up on dhcp4 when it gets dhcp6 lease
Keywords:
Status: CLOSED DUPLICATE of bug 1961666
Alias: None
Product: OpenShift Container Platform
Classification: Red Hat
Component: RHCOS
Version: 4.8
Hardware: Unspecified
OS: Unspecified
unspecified
unspecified
Target Milestone: ---
: ---
Assignee: Dusty Mabe
QA Contact: Michael Nguyen
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2021-06-03 13:57 UTC by Michael Hrivnak
Modified: 2021-06-22 20:09 UTC (History)
6 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2021-06-22 20:09:13 UTC
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)
boot log (57.53 KB, text/plain)
2021-06-03 13:57 UTC, Michael Hrivnak
no flags Details

Description Michael Hrivnak 2021-06-03 13:57:53 UTC
Created attachment 1788846 [details]
boot log

OCP Version at Install Time: quay.io/openshift-release-dev/ocp-release:4.8.0-fc.3-x86_64
RHCOS Version at Install Time: 48.84.202105062123-0
Platform: none
Architecture: x86_64


What are you trying to do? What is your use case?
Boot a minimal discovery live ISO created by assisted-installer (booting with libvirt on a RHEL 8.4 host), in order to add a worker to a cluster. libvirt is configured for bridge networking, so the VM is on my LAN.


What happened? What went wrong or what did you expect?

My local router seems to give out dhcp6 leases more quickly than dhcp4. During boot, NetworkManager appears to give up on the dhcp4 lease once it gets a dhcp6 lease. See the attached log at 21 seconds.

The boot process then tries to retrieve the rootfs from mirror.openshift.com, but because we don't support ipv6 there yet and the system only has an ipv6 address, it gets stuck in a loop of endlessly failing to GET the rootfs.


What are the steps to reproduce your issue? Please try to reduce these steps to something that can be reproduced with a single RHCOS node.

I will attach the ignition that's embedded in the minimal ISO.

I can work around this issue reliably by adding kernel argument "ipv6.disable=1".

Comment 2 Dusty Mabe 2021-06-03 20:13:29 UTC
This is a current known issue being worked on by the NetworkManager team. See https://bugzilla.redhat.com/show_bug.cgi?id=1928345#c14

Comment 3 Dusty Mabe 2021-06-03 20:15:55 UTC
In your case (since you don't need ipv6) you can also just use `ip=dchp` (see https://bugzilla.redhat.com/show_bug.cgi?id=1928345#c7).

Comment 4 Dusty Mabe 2021-06-03 20:16:22 UTC
(In reply to Dusty Mabe from comment #3)
> In your case (since you don't need ipv6) you can also just use `ip=dchp`
> (see https://bugzilla.redhat.com/show_bug.cgi?id=1928345#c7).

typo: `ip=dhcp`

Comment 5 Michael Hrivnak 2021-06-03 21:18:51 UTC
Thanks. I apparently don't have the ability to view that issue, but feel free to close this one as a duplicate if you think that makes sense.

Comment 6 Dusty Mabe 2021-06-11 20:50:14 UTC
related upstream NetworkManager issue: https://gitlab.freedesktop.org/NetworkManager/NetworkManager/-/issues/729

Comment 7 Dusty Mabe 2021-06-22 20:09:13 UTC
(In reply to Dusty Mabe from comment #2)
> This is a current known issue being worked on by the NetworkManager team.
> See https://bugzilla.redhat.com/show_bug.cgi?id=1928345#c14

That bug was closed as a duplicate of https://bugzilla.redhat.com/show_bug.cgi?id=1961666

I'll do the same here.

*** This bug has been marked as a duplicate of bug 1961666 ***


Note You need to log in before you can comment on or make changes to this bug.