Description of problem: On systems with DASD disks, the upgrade from old luks to new does not work as intended. The investigation that followed revealed that the method for upgrading LUKS devices would have to significantly change to support DASD disks and would need to be moved into the initrd (rather than the real root). Version-Release number of selected component (if applicable): How reproducible: Steps to Reproduce: 1. 2. 3. Actual results: Expected results: Additional info: specifically this will not work: https://github.com/openshift/os/blob/master/overlay.d/05rhcos/usr/libexec/rhcos-upgrade-root-filesystem-container#L39. As github issue has been created #448 https://github.com/openshift/os/issues/448
Re-assigning to Prashanth as he initially discovered the bug and created a github ticket to investigate into the problem.
Setting target release as 4.7 and blocker+ for now. We are waiting for the x86 team to investigate further as a workaround has been proposed. This bug may change to a non-blocker if the workaround is accepted and implemented
Keeping the "Blocker+" flag on until we can confirm that this bug does not prevent upgrade.
Hi Prashanth, if this bug will not be resolved before the end of this sprint, can we add "UpcomingSprint"?
Adding "UpcomingSprint" as this will not be resolved during this sprint. Will re-evaluate blocker status next week
Hi Prashanth, just checking in once again. Is this bug still a blocker. If we are still able to upgrade, then this bug shouldn't be a blocker.
After chatting with Prashanth, we think this bug should be reassigned to the RHCOS team as the fix for this bug is being worked on by their team. Hi RHCOS team, I wanted to bring up this blocker bug to your attention. Since your team is working on the "DASD disks + LUKS in RHCOS/OCP 4.7" fix, I am re-assigning this bug from Multi-Arch to your sub-component. Please re-assign as needed. Thank you.
I believe Nikita and/or Benjamin are working on a fix for this.
We've reverted the upgrade code for 4.7, and we will not reintroduce it without properly handling DASD disks.
Prashanth and I chatted and he has tested rhcos upgrade from 4.6->4.7 and it has worked fine. In addition, we would prefer for our IBM Power and Z colleagues to perform the basic testing on this too. Hi @aprabhak and @wvoesch would your teams have the cycle to test an OCP upgrade from 4.6->4.7 on DASD systems from your (P & Z) side to confirm that this bug would not prevent rhcos upgrade? We want to make sure that this bug is not blocking anything.
On the Z side we test that on a regular basis. We have tested the upgrade from from 4.6 to 4.7.0-fc.2.
On P, we tested the following paths. Upgrade from 4.6.10 --> 4.7.0-fc-2 ( libvirt environment ) Upgrade from 4.6.10 --> 4.7.0-fc-2 ( powervm environment )
Since our P & Z colleagues have tested this bug and ensured that this bug is not blocking upgrade, I am de-escalating this bug to "Blocker-" as it is still a valid bug but will not cause blocking of rhcos upgrade
The fc2 image has the removal of the upgrade scripts which should fix this issue and looks like from the comments above it has been tested. Marking it verified.
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 (Moderate: OpenShift Container Platform 4.7.0 security, bug fix, and enhancement update), 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/RHSA-2020:5633