FS#76518 - [systemd] 252.1-1 fails to boot because of swap on LVM on LUKS not found

Attached to Project: Arch Linux
Opened by solsTiCe (zebul666) - Friday, 11 November 2022, 02:49 GMT
Last edited by Toolybird (Toolybird) - Sunday, 13 November 2022, 05:45 GMT
Task Type Bug Report
Category Packages: Extra
Status Closed
Assigned To No-one
Architecture All
Severity Low
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 0
Private No

Details

My setup:
LVM in LUKS with root partition and swap as PV.

Since systemd-252.1 it fails to complete the boot because it halts at waiting for /dev/thevgname/swap device indefinitely

First, I tried to fix it by removing the resume= options in kernel parameters. But, no change
Then, commented the line about swap in /etc/fstab.

Rerun, mkinitcpio -P, just in case.
During my trials, it did work once, but I don't know why, and coulnd' reproduce later on.

I had to revert to systemd-251

I don't think this is relevant but I have to say that I am using Secure Boot, booting with systemd-boot and shim method.

everyhting is fine with systemd-251.7-4
This task depends upon

Closed by  Toolybird (Toolybird)
Sunday, 13 November 2022, 05:45 GMT
Reason for closing:  None
Additional comments about closing:  Reporter says "After finding the real issue, this is working for me now"
Comment by solsTiCe (zebul666) - Friday, 11 November 2022, 02:53 GMT
And also using linux-zen kernel, by the way, that was just upgraded too to 6.0.8-zen1
Comment by solsTiCe (zebul666) - Friday, 11 November 2022, 03:38 GMT
I never had the problem before but some other people might have had a problem that look almost the same. https://bbs.archlinux.org/viewtopic.php?pid=2034147#p2034147

I need to test those units to see if it helps. Too late for that today
Comment by vvlad (vvlad) - Friday, 11 November 2022, 09:37 GMT
Are you using busybox or systemd in the intiramfs?

I have a somewhat similar issue while using busybox. If I have the `resume` hook, it won't load some modules (in my case zfs). The resume partition is a plain dm-crypt partition and is loaded OK. `mkinitcpio` doesn't complain about anything. Reverting to systemd 251.7 fixes the issue.
Comment by solsTiCe (zebul666) - Friday, 11 November 2022, 14:02 GMT
systemd; specifically systemd, sd-encrypt and sd-vconsole

$ grep -v ^# /etc/mkinitcpio.conf |grep HOOKS
HOOKS=(base systemd autodetect modconf keyboard sd-vconsole block sd-encrypt lvm2 filesystems fsck)

$ grep -v ^# /etc/sbupdate.conf |grep CMDLINE
CMDLINE_DEFAULT="quiet loglevel=3 systemd.show_status=auto rd.udev.log_level=3 vt.global_cursor_default=0 root=/dev/a2000vg/root rd.luks.name=12345678-1234-1234-1234-123456789012=a2000luks rd.luks.options=discard,keyfile-offset=12345,keyfile-size=2048 rd.luks.key=/dev/disk/by-id/usb-hp_yyyy_XXXXXXXXXXXXX-0:0 resume=/dev/a2000vg/swap

$ grep -v ^# /etc/fstab |grep swap
UUID=1111111-1111-4444-aaaa-123456789012 none swap defaults,pri=20 0 0
/dev/mapper/swap none swap defaults,pri=10 0 0
Comment by solsTiCe (zebul666) - Saturday, 12 November 2022, 13:35 GMT
After having an issue too on my laptop with a different error, I suspected a problem with sbupdate-git [AUR]: https://github.com/andreyv/sbupdate/issues/58

It turns out the root cause is a systemd bug: https://github.com/systemd/systemd/issues/25201

For me, this is working with latest sbupdate-git, because it includes a work-around for the systemd bug(?).
Comment by solsTiCe (zebul666) - Saturday, 12 November 2022, 13:37 GMT
this is related to  FS#76468 

Loading...