Yes, it is what we verified, so from our side we can say its verified.

I agree with zabitter that we just need to check the generated install-config.yaml, and I believe that is what rh-ee-dasulin verified.

If we can verify that the additionalNTPSources data ends up in the Provisioning CR in the installed cluster then I am happy. From that point on is already tested by the metal platform team.

zabitter Should this be tested by the ABI QE team? We have verified that the generated install-config.yaml contains the NTP configuration, but it may be good to also test that ABI can actually use this to add a day 2 node.

mburman@redhat.com I believe this fix will be included in 4.19. Note that there is also a 4.18 backport:  OCPBUGS-45183.

jhernand-rh alkaplan@redhat.com Hi, pls add fix version, thanks

CPaaS Service Account mentioned this issue in merge request !689 of rh-ecosystem-edge / assisted-installer / assisted installer projects on branch rhai-1.0-rhel-8_upstream_d16fb4c1ba6bd74ed0fd92dbedc4fdcc:

Updated 3 upstream sources

Hi jhernand-rh,

Bugs should not be moved to Verified without first providing a Release Note Type("Bug Fix" or "No Doc Update") and for type "Bug Fix" the Release Note Text must also be provided. Please populate the necessary fields before moving the Bug to Verified.

Thanks zabitter . I think that the following change should address the issue then:

https://github.com/openshift/assisted-service/pull/7028

Note that I only verified that the generated install-config.yaml includes the new field, but I didn't verify that a baremetal day 2 node can be added, as I don't have the required infrastructure available. But I believe it will work, as it is the same code path.

jhernand-rh no, you still need the MachineConfigs. Don't change anything that you are already doing.

There are 3 different first-boot images for day 1 installs on platform:baremetal (Assisted, ABI, Metal Platform). Therefore, there are 3 different tickets for ensuring immediate NTP sync (MGMT-18411, OCPBUGS-39231, METAL-1022).

On day 2 when adding a node, there are the same 3 different first-boot methods, and NTP sync for them is covered by the same 3 tickets.

However! Day 1 and day 2 are orthogonal. The installer you choose on day 1 does not determine what method you use to scale out on day 2.

The bug here is that the metal platform configuration for day 2 from METAL-1022 only exists if you also used metal platform to install on day 1. We need to populate the install-config field with the data assisted-service has so that we get the same configuration regardless of install method.

For ABI purposes we'll want to backport this to 4.18. (For assisted we still have time to land it in the next ACM.) Potentially we could change the component to Installer / Agent Based Installer - since the same change will apply to (and is needed by) both, it's somewhat academic which component we use.