Project

General

Profile

Actions

tickets #162401

closed

falkor21.i.o.o freezes at POST

Added by crameleon 6 months ago. Updated about 2 months ago.

Status:
Resolved
Priority:
Normal
Assignee:
Category:
Physical infrastructure / Hardware
Target version:
-
Start date:
2024-06-17
Due date:
% Done:

0%

Estimated time:

Description

After rebooting falkor21 after installing updates, it freezes at POST before booting the OS. It does not react to any keyboard input (F11, DEL, Ctrl+Alt+Del). Power cycling the machine makes it repeat the startup process always until the same point.


Files


Related issues 2 (1 open1 closed)

Related to openSUSE admin - tickets #162326: Leap 15.6 upgrade diaryBlockedcrameleon2024-06-12

Actions
Blocked by openSUSE admin - tickets #162428: Activate Supermicro OOB licensesResolvedcrameleon2024-06-18

Actions
Actions #1

Updated by crameleon 6 months ago

  • Private changed from Yes to No
Actions #2

Updated by crameleon 6 months ago

Actions #3

Updated by crameleon 6 months ago

  • Status changed from New to In Progress
  • Assignee set to crameleon

Hitting DEL/F11 early (before the problematic message) grants entry to the EFI setup and boot menu respectively. Manually selecting the boot entry yields the same message though.
Setting up a HTTP server to host a live media to attempt booting from somewhere other than the main OS RAID.

Actions #4

Updated by crameleon 6 months ago

Cannot attach ISO file for virtual CDROM, the firmware is not activated with a "SFT-OOB-LIC" license.
Alternative is using HTTP boot, but it would require some work:

  • configure DHCP and HTTP server somewhere
  • write a GRUB configuration and place it on the HTTP server
  • reconfigure switch ports to access ports in the relevant VLAN

... a bit overkill just for testing if booting from a different media works.

I will ask about the license or for someone to plug in a pen drive.

Actions #5

Updated by crameleon 6 months ago

Actions #6

Updated by crameleon 6 months ago

Acquired the licenses, made #162428.

Actions #7

Updated by crameleon 6 months ago

Booting from live media works, so maybe the 15.6 upgrade nuked the bootable RAID afterall.

Actions #8

Updated by crameleon 6 months ago

Within the live environment, the RAID is found to be fine and so is the boot partition.
Inside a chroot, pbl --config --install is found to create an odd /boot/efi/EFI/BOOT/BOOTX64.EFI file on 15.6. On 15.5, we only have /boot/efi/EFI/opensuse/grubx64.efi. Internet research suggests this to be a fallback path for some platforms. Removing it does unfortunately not help.

Actions #9

Updated by crameleon 6 months ago

Some debugging in the underlying /usr/lib/bootloader/grub2-efi/install does not reveal anything causing this, suggesting it's a change in grub2-install. Unsure whether it's actually problematic though or if the issue is somewhere else. One idea would be to re-install grub2 using a grub package from 15.5.

Actions #10

Updated by crameleon 6 months ago ยท Edited

Indeed, after some fiddling to get networking in the live environment (the rescue ISO ships NetworkManager but our real system has wicked configuration - eventually I just manually configured it) installing the grub2 packages from 15.5

grub2-2.06-150500.29.25.12.x86_64.rpm
grub2-i386-pc-2.06-150500.29.25.12.noarch.rpm
grub2-x86_64-efi-2.06-150500.29.25.12.noarch.rpm

through the chroot and running grub2-install makes the system boot again.

Actions #11

Updated by crameleon 6 months ago

  • Status changed from In Progress to Blocked
Actions #12

Updated by crameleon 6 months ago

  • Status changed from Blocked to Workable
Actions #13

Updated by crameleon 5 months ago

  • Status changed from Workable to In Progress

The suggestion from Michael yields success.

Actions #14

Updated by crameleon 5 months ago

Michael sent a patch upstream. :)

Actions #15

Updated by crameleon 5 months ago

Remaining machines upgraded with added 25_bli workaround.

Actions #16

Updated by crameleon 4 months ago

Patch landed in Factory, waiting for 15.6 maintenance update.

Actions #17

Updated by crameleon 3 months ago

  • Status changed from In Progress to Workable

To revert the 25_bli workarounds and to test if the machines boot with the default file as shipped with the package after the update.

Actions #18

Updated by crameleon about 2 months ago

  • Status changed from Workable to Resolved

Updates installed, workarounds reverted, machines boot correctly again.

Actions

Also available in: Atom PDF