Bug 2087225 - Booting systemd 251~rc3-1.fc37 fails - probably luks support related
Summary: Booting systemd 251~rc3-1.fc37 fails - probably luks support related
Keywords:
Status: CLOSED RAWHIDE
Alias: None
Product: Fedora
Classification: Fedora
Component: systemd
Version: rawhide
Hardware: Unspecified
OS: Unspecified
unspecified
unspecified
Target Milestone: ---
Assignee: systemd-maint
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2022-05-17 15:49 UTC by Bruno Wolff III
Modified: 2022-05-25 14:25 UTC (History)
13 users (show)

Fixed In Version: systemd-251.1-1.fc37
Clone Of:
Environment:
Last Closed: 2022-05-24 22:18:18 UTC
Type: Bug
Embargoed:


Attachments (Terms of Use)
journalctl for a sample bad boot (206.74 KB, text/plain)
2022-05-17 15:49 UTC, Bruno Wolff III
no flags Details

Description Bruno Wolff III 2022-05-17 15:49:04 UTC
Created attachment 1880537 [details]
journalctl for a sample bad boot

Description of problem:
The boot hangs when using systemd 251~rc3-1.fc37. Booting works with systemd 
251~rc2-2.fc37.

Some encrypted devices aren't mounted and cryptsetup hangs when trying to do it manually to finish up the boot.

Downgrading using a live image got things working again.

How reproducible:
It happens every boot, but might not happen with all systems.

Comment 1 Zbigniew Jędrzejewski-Szmek 2022-05-22 12:35:21 UTC
https://koji.fedoraproject.org/koji/taskinfo?taskID=87360713 has a scratch builds with one patch reverted.
Could you test if it fixes the issue?

Comment 2 Martin Wilck 2022-05-23 16:16:32 UTC
Could you describe your storage layout? Is the device for /boot (part of) the LUKS device?

Also, please attach a boot log taken with "systemd.log_level=debug".

One possible issue is lvm2-monitor.service which is obviously started too early ("WARNING: Device /dev/sda1 not  initialized in udev database even after waiting 10000000 microseconds.").

Comment 3 Bruno Wolff III 2022-05-23 16:59:46 UTC
/boot is not encrypted.
I'll test things sometime today. I have the 251-1 release installed right now, but haven't tried rebooting yet. I'll 
try it at a convenient point today. If that doesn't work, I'll try the scratch build you pointed to.

Comment 4 Bruno Wolff III 2022-05-23 23:25:23 UTC
The 251-1 release still has the problem. I'm about to test the scratch build.

Comment 5 Bruno Wolff III 2022-05-23 23:41:47 UTC
The scratch build worked for me.

Comment 6 Zbigniew Jędrzejewski-Szmek 2022-05-24 22:18:18 UTC
This should be fixed, but please test.

Comment 7 Bruno Wolff III 2022-05-24 22:41:07 UTC
I tested systemd-251.1-1.fc37.x86_64 and it worked fine. Thanks for fixing this.

Comment 8 Martin Wilck 2022-05-25 14:25:37 UTC
The fix that Zbigniew provided simply reverted a patch from v251. That's fine, but meanwhile I've come up with a patch that fixes the LUKS issue (for me, at least) without re-introducing the breakage in multipath setups that v250 and earlier releases suffered from.

Therefore, if you can, please test with the code in https://github.com/systemd/systemd/pull/23489. Unfortunately I don't know how to provide test binaries for Fedora, any help would be appreciated.


Note You need to log in before you can comment on or make changes to this bug.