TripleO deployment over IPV6

Julia Kreger juliaashleykreger at gmail.com
Thu Jul 27 13:29:28 UTC 2023


On Thu, Jul 27, 2023 at 4:41 AM Kushagr Gupta <kushagrguptasps.mun at gmail.com>
wrote:

> Hi Team,Julia,
>
> @Julia Thank you for the response.
> >>Additionally is this UEFI mode or Legacy BIOS mode?
> We are using UEFI.
>
> We tried introspection for two nodes from the undercloud node. Te
> following is our observation:
> 1. Undercloud is assigning an IPV6 IP to the node.
>
> The PXE boot Failed with the following error:
> PXE-E99: Unexpected network error.
>
>
I guess what is weird in this entire thing is it sounds like you're
shifting over to what appears to be OPROM boot code in a network interface
card, which might not support v6. Then again a port mirrored packet capture
would be the needful item to troubleshoot further.


> On further debuging this issue, We found that a container: ironic_pxe_tftp
> is in unhealthy state. But this container should be running.
> On checking the logs, we  found the following:
> "
> [root at undercloud-ete log]# podman logs 356715c49cd4
>
> dnsmasq: bad command line options: try --help
>
> dnsmasq: bad command line options: try --help
>
> dnsmasq: bad command line options: try --help
>
>
Are you able to extract the exact command line which is being passed to the
dnsmasq process for that container launch?

I guess I'm worried if somehow dnsmasq changed or if an old version is
somehow in the container image you're using.

"
> We had previously raised this issue in 2022 as well:
>
> https://lists.openstack.org/pipermail/openstack-discuss/2022-June/029062.html
>
> For this a bug was raised which was fixed
> https://bugs.launchpad.net/tripleo/+bug/1978892
>
> We cross-verified, and the changes are present on our undercloud. But
> still we are unable to start the ironic_pxe_tftp container.
>
> Could you please help us with this?
>
> Thanks and Regards
> Kushagra Gupta
>
> On Tue, Jul 25, 2023 at 8:10 PM Julia Kreger <juliaashleykreger at gmail.com>
> wrote:
>
>> Greetings,
>>
>> I don't think I saw your original email. I suspect because you indicated
>> there was an attachment, it might have been filtered for moderation by the
>> mailing list.
>>
>>
>> On Tue, Jul 25, 2023 at 6:24 AM Lokendra Rathour <
>> lokendrarathour at gmail.com> wrote:
>>
>>> Hi Takashi/ Team,
>>> Any inputs with respect to the same will be of great help.
>>>
>>> Thanks,
>>> Lokendra
>>>
>>>
>>> On Tue, Jul 25, 2023 at 4:29 PM Kushagr Gupta <
>>> kushagrguptasps.mun at gmail.com> wrote:
>>>
>>>> Hi Team,
>>>>
>>>> We were trying to perform IPV6-based deployment of TripleO.
>>>> We are following the link:
>>>>
>>>> https://access.redhat.com/documentation/en-us/red_hat_openstack_platform/16.0/html/bare_metal_provisioning/sect-deploy#sect-example-templates
>>>>
>>>>
>>>>
>>>
>> Try following the more recent documents, specifically OSP16.1 or 16.2.
>>
>>> As per our environment, we set the "ipv6_address_mode" to
>>>> "dhcpv6-stateful" and tried undercloud deployment. It was successful.
>>>>
>>>>
>>>>
>>> Good, that is required for IPv6 PXE.
>>
>>> But when we try node introspection, openstack is not providing an IP
>>>> that will be used for PXE booting the system. We have enabled PXE booting
>>>> on the IPV6 network.
>>>> On Dell machines, we get the following error:
>>>> "PXE-E16: No offer received."
>>>>
>>>
>> I suspect we will need to see the configuration. Paste.opendev.org?
>>
>>> For your reference, the undercloud file which we used is attached.
>>>> We also tried to find the communication happening on the br-ctlplane
>>>> interface when the machine tries to boot over IPV6 PXE boot. I am attaching
>>>> those tcpdump records as well. We are unable to find the MAC address of our
>>>> machines.
>>>>
>>> I suspect this might be in the territory of the hardware, while being
>> set for IPv6 network booting, maybe the configuration needs to be
>> checked/reasserted on the interface level. It sounds as if it is using a
>> different interface.
>>
>> Not seeing the expected MAC addresses is rather suspect in this case,
>> although PXE with v6 is a little different pattern wise. Have you tried
>> mirroring the port your expecting to see traffic cross from the physical
>> machine to see if it is seeing the announcements and issuing the DHCPv6
>> packets required?
>>
>> Additionally is this UEFI mode or Legacy BIOS mode?
>>
>>> *Questions:*
>>>> 1. Is it even possible to perform tripleO deployment with IPV6?
>>>> 2. Do we need to do some extra settings, so that our baremetal machines
>>>> can be PXE booted?
>>>>
>>>> Thanks and Regards
>>>>
>>>
>>>
>>> --
>>> ~ Lokendra
>>> skype: lokendrarathour
>>>
>>>
>>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.openstack.org/pipermail/openstack-discuss/attachments/20230727/345000c7/attachment-0001.htm>


More information about the openstack-discuss mailing list