action #69406
closed[y][timeboxed:12h] Fix console switching on PowerVM during installation
0%
Description
Motivation¶
Validation for encrypted disk activation was added in scope of poo#68962. The module validate_activation_encrypted_partition was excluded for PowerVM, because while switching to install-shell console, black screen was appeared instead of the right tty console. The module should be enabled for PowerVM as well, so that the issue with the console switching need to be investigated and fixed.
What we need to check what is different in installation/logs_from_installation_system as console switching works fine there.
Acceptance criteria¶
validate_activation_encrypted_partitionis added toschedule/yast/encryption/cryptlvm+activate_existing_spvm.yaml;- The
validate_activation_encrypted_partitionmodule is passed on PowerVM.
Updated by riafarov over 3 years ago
- Subject changed from [y] Fix console switching on PowerVM during installation to [y][timeboxed:12h] Fix console switching on PowerVM during installation
- Description updated (diff)
- Category changed from Refactor/Code Improvements to Spike/Research
- Status changed from New to Workable
Updated by JRivrain over 3 years ago
- Status changed from Workable to In Progress
Updated by JRivrain over 3 years ago
We just need to use use_ssh_serial_console, then switching works back and forth: https://openqa.suse.de/tests/4529101#step/validate_activation_encrypted_partition/2. Unfortunately the test fails as the device shows as "active" instead of "active and in use" : https://openqa.suse.de/tests/4529102#step/validate_activation_encrypted_partition/6. Trying to figure out what is wrong.
Updated by JRivrain over 3 years ago
- Status changed from In Progress to Feedback
So as @ybonatakis pointed out, the problem probably comes from this: https://gitlab.com/cryptsetup/cryptsetup/-/blob/master/src/cryptsetup.c#L797
my guess is that cryptsetup is not done crypting the disk from that step https://openqa.suse.de/tests/4536402#step/bootloader_start/37 so if I'm right, just putting the module later may do the trick. However I cannot currently test this due to an outage.
PR that may be used: https://github.com/os-autoinst/os-autoinst-distri-opensuse/pull/10786/files
Putting this in feedback, please re-assign or resolve as I'll be away for a week.
Updated by riafarov over 3 years ago
- Status changed from Feedback to Blocked
This is blocked now due to overheating in the server room, so machines are staying off still. We use #69754 for tracking, I will add this ticket to related, so we unblock it once servers are running again.
Updated by riafarov over 3 years ago
- Related to action #69754: [y][u] tests fail in bootloader_start for ppc64le - PowerVM workers not available added
Updated by riafarov over 3 years ago
- Due date changed from 2020-08-11 to 2020-08-25
Updated by riafarov over 3 years ago
- Due date changed from 2020-08-25 to 2020-09-22
Updated by riafarov over 3 years ago
- Due date changed from 2020-09-22 to 2020-10-06
- Status changed from Blocked to Workable
- Assignee deleted (
JRivrain)
Machines are up and running.
Updated by riafarov over 3 years ago
- Project changed from openQA Tests to qe-yam
- Category deleted (
Spike/Research)
Updated by riafarov over 3 years ago
- Status changed from Workable to In Progress
Updated by riafarov over 3 years ago
So switching from root console to user and back works fine: https://openqa.suse.de/tests/4940815#
No need for the change, checking methods to use serial console.
Updated by riafarov over 3 years ago
- Status changed from In Progress to Feedback
https://github.com/os-autoinst/os-autoinst-distri-opensuse/pull/11350
So what didn't work is switching to install-shell second time, I've fixed that and removed special handling of powerVM for those scenarios.
log-console would not work either, but we don't have cases where it would be called twice without reset of the consoles, but still fixed too.