Bug 15773

Summary: AB-INT: psplash: AssertionError: 3 != 0 : SYSTEMD_BUS_TIMEOUT=240s systemctl status --full --failed
Product: [QA/Testing] Runtime Testing Reporter: Mathieu Dubois-Briand <mathieu.dubois-briand>
Component: generalAssignee: Mikko Rapeli <mikko.rapeli>
Status: RESOLVED FIXED QA Contact:
Severity: normal    
Priority: Medium+ CC: randy.macleod, yi.zhao
Version: unspecified   
Target Milestone: 5.2 M3   
Hardware: x86   
OS: Multiple   
Whiteboard: AB-INT
OS type for building Yocto: --- Type of Regression: ---
Verified: Documentation change: No (bug/feature does not impact docs)

Description Mathieu Dubois-Briand 2025-02-25 12:44:04 UTC
We see this AB-INT quite often for a week or two:

    304 Traceback (most recent call last):
    305   File "/srv/pokybuild/yocto-worker/meta-arm/build/meta/lib/oeqa/core/decorator/__init__.py", line 35, in wrapped_f
    306     return func(*args, **kwargs)
    307            ^^^^^^^^^^^^^^^^^^^^^
    308   File "/srv/pokybuild/yocto-worker/meta-arm/build/meta/lib/oeqa/runtime/cases/systemd.py", line 100, in test_systemd_failed
    309     output += self.systemctl('status --full --failed')
    310               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    311   File "/srv/pokybuild/yocto-worker/meta-arm/build/meta/lib/oeqa/runtime/cases/systemd.py", line 26, in systemctl
    312     self.assertEqual(status, expected, message)
    313 AssertionError: 3 != 0 : SYSTEMD_BUS_TIMEOUT=240s systemctl status --full --failed 
    314 x psplash-systemd.service - Start psplash-systemd progress communication helper
    315      Loaded: loaded (/usr/lib/systemd/system/psplash-systemd.service; static)
    316      Active: failed (Result: exit-code) since Mon 2025-02-24 01:28:17 UTC; 15min ago
    317    Duration: 531ms
    318  Invocation: 463bbdc34ab1465495d76e402819f06a
    319    Main PID: 262 (code=exited, status=1/FAILURE)
    320    Mem peak: 1.1M
    321         CPU: 165ms
    322 
    323 Feb 24 01:28:17 sbsa-ref systemd[1]: Started Start psplash-systemd progress communication helper.
    324 Feb 24 01:28:17 sbsa-ref psplash-systemd[262]: Error unable to open fifo
    325 Feb 24 01:28:17 sbsa-ref systemd[1]: psplash-systemd.service: Main process exited, code=exited, status=1/FAILURE
    326 Feb 24 01:28:17 sbsa-ref systemd[1]: psplash-systemd.service: Failed with result 'exit-code'.
Comment 1 Mathieu Dubois-Briand 2025-02-25 12:47:22 UTC
meta-arm debian12-vk-5 mathieu/master-next completed at 2025-02-18T14:22:14Z
https://autobuilder.yoctoproject.org/valkyrie/#/builders/75/builds/986/steps/22/logs/stdio

meta-arm opensuse156-vk-1 master-next completed at 2025-02-18T16:14:46Z
https://autobuilder.yoctoproject.org/valkyrie/#/builders/75/builds/991/steps/23/logs/stdio

meta-arm ubuntu2204-vk-3 mathieu/master-next completed at 2025-02-18T18:28:41Z
https://autobuilder.yoctoproject.org/valkyrie/#/builders/75/builds/992/steps/22/logs/stdio

meta-arm ubuntu2204-vk-3 master completed at 2025-02-19T01:42:09Z
https://autobuilder.yoctoproject.org/valkyrie/#/builders/75/builds/994/steps/22/logs/stdio

meta-arm alma9-vk-1 master-next completed at 2025-02-19T10:42:12Z
https://autobuilder.yoctoproject.org/valkyrie/#/builders/75/builds/997/steps/22/logs/stdio

meta-arm fedora39-vk-2 master completed at 2025-02-22T01:50:42Z
https://autobuilder.yoctoproject.org/valkyrie/#/builders/75/builds/1017/steps/22/logs/stdio

meta-arm fedora40-vk-2 master completed at 2025-02-24T01:45:23Z
https://autobuilder.yoctoproject.org/valkyrie/#/builders/75/builds/1021/steps/22/logs/stdio
Comment 2 Mathieu Dubois-Briand 2025-02-27 12:31:18 UTC
meta-arm alma8-vk-2 master-next completed at 2025-02-25T17:39:30Z
https://autobuilder.yoctoproject.org/valkyrie/#/builders/75/builds/1036/steps/23/logs/stdio

meta-arm stream9-vk-1 master completed at 2025-02-27T01:53:58Z
https://autobuilder.yoctoproject.org/valkyrie/#/builders/75/builds/1053/steps/22/logs/stdio

meta-arm debian11-vk-2 mathieu/master-next completed at 2025-02-27T10:54:31Z
https://autobuilder.yoctoproject.org/valkyrie/#/builders/75/builds/1054/steps/22/logs/stdio
Comment 3 Randy MacLeod 2025-02-27 15:44:14 UTC
Patch on the list.
Comment 4 Mathieu Dubois-Briand 2025-03-03 16:06:46 UTC
meta-arm stream9-vk-1 mathieu/master-next completed at 2025-02-27T18:13:15Z
https://autobuilder.yoctoproject.org/valkyrie/#/builders/75/builds/1058/steps/22/logs/stdio
Comment 5 Randy MacLeod 2025-03-06 15:58:52 UTC
https://git.openembedded.org/openembedded-core/commit/?id=580ae81e102bf999cb89f05430c737210253d90a

helps but there's more work to do.
Comment 6 Mikko Rapeli 2025-03-10 07:33:54 UTC
Is this bug still happening in CI builds and tests with qemu?
Comment 7 Mathieu Dubois-Briand 2025-03-10 09:38:02 UTC
It looks like we didn't get the issue for one week now.
Comment 8 Mikko Rapeli 2025-03-10 10:11:43 UTC
(In reply to Mathieu Dubois-Briand from comment #7)
> It looks like we didn't get the issue for one week now.

I think this is because the previous failing build was missing
https://git.openembedded.org/openembedded-core/commit/?id=580ae81e102bf999cb89f05430c737210253d90a
which was applied Feb 28th and the build ran Feb 27th without it.

Would be nice to confirm this but I can't check this from the logs or the non-public commits.

Did

meta                 
meta-poky            
meta-yocto-bsp       = "mathieu/master-next:ffe27fc69ff0de826e3d8dbffaa6e099c2d6b588"

contain the fix above or not?

If not, then I think this is fixed now.
Comment 9 Mathieu Dubois-Briand 2025-03-10 10:52:57 UTC
No I believe it did not. Branch of the said build can be found at https://git.yoctoproject.org/poky-ci-archive/log/?h=autobuilder.yoctoproject.org/valkyrie/a-full-1097 .

So this is probably fixed.
Comment 10 Mikko Rapeli 2025-03-11 07:00:17 UTC
Marking as fixed then.
Comment 11 Mathieu Dubois-Briand 2025-05-12 13:41:38 UTC
I got this one on master, yet I believe the fix is still present. Only change I
can see so far is a psplash update.

meta-arm debian12-vk-5 master completed at 2025-05-12T01:26:07Z
https://autobuilder.yoctoproject.org/valkyrie/#/builders/75/builds/1509/steps/22/logs/stdio
Comment 12 Mikko Rapeli 2025-05-12 14:02:21 UTC
https://git.yoctoproject.org/meta-arm/commit/?id=a86f62f1449ce0e341ccf1a1004a1177e4db4e9d
and
https://git.yoctoproject.org/meta-arm/commit/?id=45daeba052758831e5b589e6fab9f5a040cc6d37

show that meta-arm sbsa machine has had framebuffer related issues since the beginning. So psplash may hit this too.
Comment 13 Mathieu Dubois-Briand 2025-08-21 06:55:11 UTC
I had this one last week. Waited a bit to see if it reproduces, but it looks
like it is a loner. Not sure if it is the intermittent issue coming back or just
something that was bad in my build.

meta-arm fedora40-vk-3 mathieu/master-next completed at 2025-08-13 20:59:10+00:00
https://autobuilder.yoctoproject.org/valkyrie/#/builders/75/builds/2097/steps/22/logs/stdio
Comment 14 Mikko Rapeli 2025-08-21 07:08:27 UTC
(In reply to Mathieu Dubois-Briand from comment #13)
> I had this one last week. Waited a bit to see if it reproduces, but it looks
> like it is a loner. Not sure if it is the intermittent issue coming back or
> just
> something that was bad in my build.
> 
> meta-arm fedora40-vk-3 mathieu/master-next completed at 2025-08-13
> 20:59:10+00:00
> https://autobuilder.yoctoproject.org/valkyrie/#/builders/75/builds/2097/
> steps/22/logs/stdio

Traceback (most recent call last):
  File "/srv/pokybuild/yocto-worker/meta-arm/build/meta/lib/oeqa/core/decorator/__init__.py", line 35, in wrapped_f
    return func(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^
  File "/srv/pokybuild/yocto-worker/meta-arm/build/meta/lib/oeqa/runtime/cases/systemd.py", line 100, in test_systemd_failed
    output += self.systemctl('status --full --failed')
              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/srv/pokybuild/yocto-worker/meta-arm/build/meta/lib/oeqa/runtime/cases/systemd.py", line 26, in systemctl
    self.assertEqual(status, expected, message)
AssertionError: 3 != 0 : SYSTEMD_BUS_TIMEOUT=240s systemctl status --full --failed 
x psplash-start@fb0.service - Start psplash boot splash screen
     Loaded: loaded (/usr/lib/systemd/system/psplash-start@.service; static)
     Active: failed (Result: timeout) since Wed 2025-08-13 20:56:27 UTC; 1min 54s ago
 Invocation: 7508512f9cb84905bd737ad9c91a6c39
    Process: 263 ExecStart=/usr/bin/psplash (code=killed, signal=TERM)
   Main PID: 263 (code=killed, signal=TERM)
   Mem peak: 1.1M
        CPU: 56ms
Aug 13 20:52:27 sbsa-ref systemd[1]: Starting Start psplash boot splash screen...
Aug 13 20:56:27 sbsa-ref systemd[1]: psplash-start@fb0.service: start operation timed out. Terminating.
Aug 13 20:56:27 sbsa-ref systemd[1]: psplash-start@fb0.service: Failed with result 'timeout'.
Aug 13 20:56:27 sbsa-ref systemd[1]: Failed to start Start psplash boot splash screen.

Would be nice to see the kernel dmesg logs from the device. There could be framebuffer related errors there. Not sure but the communication with qemu framebuffer device all the way to the graphics stack of the host system so I guess there can be delays which the qemu machine SW simply can't deal with.
Comment 15 Mathieu Dubois-Briand 2025-11-27 09:23:41 UTC
Hum, again :(

meta-arm alma8-vk-1 master-next completed at 2025-11-25 15:55:21+00:00
https://autobuilder.yoctoproject.org/valkyrie/#/builders/75/builds/2655/steps/24/logs/stdio
Comment 16 Mikko Rapeli 2025-11-27 09:34:56 UTC
Traceback (most recent call last):
  File "/srv/pokybuild/yocto-worker/meta-arm/build/layers/openembedded-core/meta/lib/oeqa/core/decorator/__init__.py", line 35, in wrapped_f
    return func(*args, **kwargs)
  File "/srv/pokybuild/yocto-worker/meta-arm/build/layers/openembedded-core/meta/lib/oeqa/runtime/cases/systemd.py", line 100, in test_systemd_failed
    output += self.systemctl('status --full --failed')
              ~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/srv/pokybuild/yocto-worker/meta-arm/build/layers/openembedded-core/meta/lib/oeqa/runtime/cases/systemd.py", line 26, in systemctl
    self.assertEqual(status, expected, message)
    ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError: 3 != 0 : SYSTEMD_BUS_TIMEOUT=240s systemctl status --full --failed 
x psplash-start@fb0.service - Start psplash boot splash screen
     Loaded: loaded (/usr/lib/systemd/system/psplash-start@.service; static)
     Active: failed (Result: timeout) since Tue 2025-11-25 15:50:27 UTC; 4min 2s ago
 Invocation: 6518720d00f74bf7800994ff3646e4f3
    Process: 265 ExecStart=/usr/bin/psplash (code=killed, signal=TERM)
   Main PID: 265 (code=killed, signal=TERM)
   Mem peak: 1.1M
        CPU: 99ms
Nov 25 15:46:27 sbsa-ref systemd[1]: Starting Start psplash boot splash screen...
Nov 25 15:50:27 sbsa-ref systemd[1]: psplash-start@fb0.service: start operation timed out. Terminating.
Nov 25 15:50:27 sbsa-ref systemd[1]: psplash-start@fb0.service: Failed with result 'timeout'.
Nov 25 15:50:27 sbsa-ref systemd[1]: Failed to start Start psplash boot splash screen.
Startup finished in 4.233s (firmware) + 5.340s (loader) + 3.030s (kernel) + 4min 17.571s (userspace) = 4min 30.175s.
Target boot time 270.175 exceeds systemd's TimeoutStartSec 90

This boot time feels quite slow. Was the machine really loaded? How frequently is this being hit? I presume this is pretty rare.

We could try to trace the psplash syscalls a bit more to see what is hanging for so long but I suspect it's just one of the framebuffer operations which takes for ever. This then can be caused by qemu and even the host system graphics stack which is highly loaded and slow.

What server host system was running this? Is it possible to know what kind of load the system had at the time when this happened?