Bug 1961784

Summary:	[OSP16.2]Overcloud deployment failed for ironic in overcloud UEFI boot mode configuration on partition images
Product:	Red Hat OpenStack	Reporter:	Jason Paroly <jparoly>
Component:	openstack-ironic-python-agent	Assignee:	Steve Baker <sbaker>
Status:	CLOSED ERRATA	QA Contact:	David Rosenfeld <drosenfe>
Severity:	urgent	Docs Contact:
Priority:	urgent
Version:	16.2 (Train)	CC:	amcleod, bfournie, jlabarre, mburns, nlevinki, pbabbar, pweeks, sbaker, spower
Target Milestone:	beta	Keywords:	Documentation, Regression, Triaged
Target Release:	16.2 (Train on RHEL 8.4)
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:
Fixed In Version:	openstack-ironic-python-agent-5.0.5-2.20210602014810.e3da253.el8ost	Doc Type:	Release Note
Doc Text:	In this release, EFI bootloader assets for whole-disk images are preserved during deployment, so the shim bootloader is no longer overwritten. This ensures that Secure Boot is switched on after deployment.	Story Points:	---
Clone Of:		Environment:
Last Closed:	2021-09-15 07:15:04 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 Jason Paroly 2021-05-18 17:18:19 UTC

Description of problem:
16.2 deployment failed for ironic in overcloud

CREATE_FAILED  ResourceInError: resources.NovaCompute: Went to status ERROR due to "Message: Exceeded maximum number of retries. Exhausted all hosts available for retrying build failures for instance bddad038-d3f1-4552-9416-d6563d54fa10., Code: 500"


Version-Release number of selected component (if applicable):
16.2


How reproducible:
everytime


Steps to Reproduce:
1. deploy the Overcloud with ironic in overcloud enabled
2.
3.

Actual results:
OC is failing


Expected results:
OC should pass


Additional info:

Comment 2 Steve Baker 2021-05-25 19:47:32 UTC

grub2 has the change which forbids it being installed to the UEFI boot path in rhel-8.4. So we need to update the documentation that UEFI boot with partition images is not supported. Section 7.2[1] needs to be updated to make it clear that if UEFI boot is enabled they will need to follow the instructions in Chapter 24[2] to build their own whole-disk image.

[1] https://access.redhat.com/documentation/en-us/red_hat_openstack_platform/16.1/html-single/director_installation_and_usage/index#setting-uefi-boot-mode
[2] https://access.redhat.com/documentation/en-us/red_hat_openstack_platform/16.1/html-single/director_installation_and_usage/index#creating-whole-disk-images

Comment 3 Steve Baker 2021-05-25 19:59:42 UTC

I'm going to look at overcloud-full preinstalling the shim and signed grub, and IPA checking for efi paths before deciding whether to run grub2-install and where to place grub.cfg

Comment 4 Steve Baker 2021-05-27 02:29:28 UTC

I just deployed a partition image overcloud-full using master IPA, and it looks like the _preserve_efi_assets path leads to a successful deployment because grub2-install is not run. 

_preserve_efi_assets was backported to ussuri IPA in the last backport push, so in theory it is a single backport to bring this to train. I think doing this will fix this CI job and unblock us for 16.2.

With images that have a populated /boot/efi the only remaining case which results in a grub2-install call is whole-disk+softraid. Train IPA doesn't support softraid at all so unless we're going to start supporting this I think nothing needs to be backported for this scenario.

Comment 5 spower 2021-06-02 14:20:37 UTC

Blocker+ flag given.

Comment 16 errata-xmlrpc 2021-09-15 07:15:04 UTC

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory (Red Hat OpenStack Platform (RHOSP) 16.2 enhancement advisory), and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHEA-2021:3483