Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.debian.bugs.dist > #1251418 > unrolled thread

Bug#1108498: vmdb2: images build on trixie for buster fail to boot

Back to article view | Back to linux.debian.bugs.dist

Contents

  Bug#1108498: vmdb2: images build on trixie for buster fail to boot Paul Gevers <elbrus@debian.org> - 2025-06-29 21:10 +0200
    Bug#1108498: vmdb2: images build on trixie for buster fail to boot Gunnar Wolf <gwolf@debian.org> - 2025-07-18 12:30 +0200
    Bug#1108498: vmdb2: images build on trixie for buster fail to boot Gunnar Wolf <gwolf@debian.org> - 2025-08-04 22:40 +0200
      Bug#1108498: vmdb2: images build on trixie for buster fail to boot Lars Wirzenius <liw@liw.fi> - 2025-08-05 11:00 +0200

#1251418 — Bug#1108498: vmdb2: images build on trixie for buster fail to boot

[Multipart message — attachments visible in raw view] — view raw

Package: vmdb2
Version: 0.40-2
Severity: normal

Hi,

I think the situation is similar to what happened during the bookworm 
release in bug #1031364, but now with the option orphan_file. Today I 
noticed that on ci.d.n the tests that use qemu fail when running tests 
on oldstable (buster) because the image (which is built with 
autopkgtest-build-qemu which uses vmdb2 under the hood) doesn't boot. 
When I run the test manually with the --show-boot option enabled in 
autopkgtest's qemu backend, the test ends like below. The test hangs for 
1 minute after """/dev/vda1 has unsupported feature(s): FEATURE_C12"""

This seems to have started to happen after we upgraded the host from 
bookworm to trixie.

I suspect that the code here needs to be expanded: 
https://sources.debian.org/src/vmdb2/0.40-2/vmdb/plugins/mkfs_plugin.py/#L63

I'll also note that it seems that code should depend on the target suite 
and should only be conditionally applied.

Paul

[    1.363469] virtio_blk virtio4: [vda] 52428800 512-byte logical 
blocks (26.8 GB/25.0 GiB)
[    1.365695] vda: detected capacity change from 0 to 26843545600
[    1.367310] printk: console [hvc0] enabled
[    1.368239]  vda: vda1
[    1.522930] ata2.00: ATAPI: QEMU DVD-ROM, 2.5+, max UDMA/100
[    1.525476] scsi 1:0:0:0: CD-ROM            QEMU     QEMU DVD-ROM 
2.5+ PQ: 0 ANSI: 5
[    1.566761] sr 1:0:0:0: [sr0] scsi3-mmc drive: 4x/4x cd/rw xa/form2 tray
[    1.568222] cdrom: Uniform CD-ROM driver Revision: 3.20
Begin: Loading essential drivers ... done.
Begin: Running /scripts/init-premount ... done.
Begin: Mounting root file system ... Begin: Running /scripts/local-top 
... done.
Begin: Running /scripts/local-premount ... done.
Begin: Will now check root file system ... fsck from util-linux 2.36.1
[/sbin/fsck.ext4 (1) -- /dev/vda1] fsck.ext4 -a -C0 /dev/vda1
/dev/vda1 has unsupported feature(s): FEATURE_C12
e2qemu-system-x86_64: terminating on signal 15 from pid 3041123 
(/usr/bin/python3)
fsck: Get a newer version of e2fsck!

/dev/vda1: ********** WARNING: Filesystem still has errors **********

fsck exited with status code 12
done.
Failure: File system check of the root filesystem failed
The root filesystem on /dev/vda1 requires a manual fsck


BusyBox v1.30.1 (Debian 1:1.30.1-6+b3) built-in shell (ash)
Enter 'help' for a list of built-in commands.

(initramfs) <VirtSubproc>: failure: timed out waiting for 'login prompt 
on serial console'
autopkgtest [06:42:37]: ERROR: testbed failure: unexpected eof from the 
testbed
Exception ignored in: <_io.TextIOWrapper name=6 encoding='UTF-8'>
BrokenPipeError: [Errno 32] Broken pipe

[toc] | [next] | [standalone]

#1253144

[Multipart message — attachments visible in raw view] — view raw

tags 1108498 + upstream
thanks

Hello Paul,

Thanks for this report. I am forward it to the upstream author.

Lars, this bug is related to filesystem options, in a very similar fashion
to #1031364 which was fixed in late 2022. Could you also please take a look
and give us your opinion?

Thanks!

Paul Gevers dijo [Sun, Jun 29, 2025 at 09:00:49PM +0200]:
>Package: vmdb2
>Version: 0.40-2
>Severity: normal
>
>Hi,
>
>I think the situation is similar to what happened during the 
>bookworm release in bug #1031364, but now with the option 
>orphan_file. Today I noticed that on ci.d.n the tests that use qemu 
>fail when running tests on oldstable (buster) because the image 
>(which is built with autopkgtest-build-qemu which uses vmdb2 under 
>the hood) doesn't boot. When I run the test manually with the 
>--show-boot option enabled in autopkgtest's qemu backend, the test 
>ends like below. The test hangs for 1 minute after """/dev/vda1 has 
>unsupported feature(s): FEATURE_C12"""
>
>This seems to have started to happen after we upgraded the host from 
>bookworm to trixie.
>
>I suspect that the code here needs to be expanded: https://sources.debian.org/src/vmdb2/0.40-2/vmdb/plugins/mkfs_plugin.py/#L63
>
>I'll also note that it seems that code should depend on the target 
>suite and should only be conditionally applied.
>
>Paul
>
>[    1.363469] virtio_blk virtio4: [vda] 52428800 512-byte logical 
>blocks (26.8 GB/25.0 GiB)
>[    1.365695] vda: detected capacity change from 0 to 26843545600
>[    1.367310] printk: console [hvc0] enabled
>[    1.368239]  vda: vda1
>[    1.522930] ata2.00: ATAPI: QEMU DVD-ROM, 2.5+, max UDMA/100
>[    1.525476] scsi 1:0:0:0: CD-ROM            QEMU     QEMU DVD-ROM 
>2.5+ PQ: 0 ANSI: 5
>[    1.566761] sr 1:0:0:0: [sr0] scsi3-mmc drive: 4x/4x cd/rw xa/form2 tray
>[    1.568222] cdrom: Uniform CD-ROM driver Revision: 3.20
>Begin: Loading essential drivers ... done.
>Begin: Running /scripts/init-premount ... done.
>Begin: Mounting root file system ... Begin: Running 
>/scripts/local-top ... done.
>Begin: Running /scripts/local-premount ... done.
>Begin: Will now check root file system ... fsck from util-linux 2.36.1
>[/sbin/fsck.ext4 (1) -- /dev/vda1] fsck.ext4 -a -C0 /dev/vda1
>/dev/vda1 has unsupported feature(s): FEATURE_C12
>e2qemu-system-x86_64: terminating on signal 15 from pid 3041123 
>(/usr/bin/python3)
>fsck: Get a newer version of e2fsck!
>
>/dev/vda1: ********** WARNING: Filesystem still has errors **********
>
>fsck exited with status code 12
>done.
>Failure: File system check of the root filesystem failed
>The root filesystem on /dev/vda1 requires a manual fsck
>
>
>BusyBox v1.30.1 (Debian 1:1.30.1-6+b3) built-in shell (ash)
>Enter 'help' for a list of built-in commands.
>
>(initramfs) <VirtSubproc>: failure: timed out waiting for 'login 
>prompt on serial console'
>autopkgtest [06:42:37]: ERROR: testbed failure: unexpected eof from 
>the testbed
>Exception ignored in: <_io.TextIOWrapper name=6 encoding='UTF-8'>
>BrokenPipeError: [Errno 32] Broken pipe




-- 

[toc] | [prev] | [next] | [standalone]

#1254973

[Multipart message — attachments visible in raw view] — view raw

tags 1108498 + patch
thanks

Hi,

I have proposed the upstream author two very simple patches, one that
unconditionally disables the use of orphan_file:

     https://gitlab.com/larswirzenius/vmdb2/-/merge_requests/155/diffs

The other one introduces a control in the mkfs plugin, by which we can
specify to vmdb that we want to build targetting an old system
("target_old"), and if so, it will avoid all newer filesystem options that
might break older systems:

     https://gitlab.com/larswirzenius/vmdb2/-/merge_requests/156/diffs

The disadvantage of this last patch is that existing vmdb2 recipes might
have to be updated adding the "target_old" option, but resulting systems
will have a (marginally) better filesystem.

Lars, please share your thoughts on this! I understand the project is in
"selfish maintenance state", but if possible, I prefer to avoid introducing
distro-specific patches.

Greetings!

    – Gunnar.

[toc] | [prev] | [next] | [standalone]

#1255007

As upstream I've merged both patches. Thani you very much, Gunnar, for
fixing and bug reporters for reporting.

On Mon, 2025-08-04 at 14:31 -0600, Gunnar Wolf wrote:
> tags 1108498 + patch
> thanks
> 
> Hi,
> 
> I have proposed the upstream author two very simple patches, one that
> unconditionally disables the use of orphan_file:
> 
>     
> https://gitlab.com/larswirzenius/vmdb2/-/merge_requests/155/diffs
> 
> The other one introduces a control in the mkfs plugin, by which we
> can
> specify to vmdb that we want to build targetting an old system
> ("target_old"), and if so, it will avoid all newer filesystem options
> that
> might break older systems:
> 
>     
> https://gitlab.com/larswirzenius/vmdb2/-/merge_requests/156/diffs
> 
> The disadvantage of this last patch is that existing vmdb2 recipes
> might
> have to be updated adding the "target_old" option, but resulting
> systems
> will have a (marginally) better filesystem.
> 
> Lars, please share your thoughts on this! I understand the project is
> in
> "selfish maintenance state", but if possible, I prefer to avoid
> introducing
> distro-specific patches.
> 
> Greetings!
> 
>     – Gunnar.

[toc] | [prev] | [standalone]

Back to top | Article view | linux.debian.bugs.dist

csiph-web