On 12/06/2020 15:05, David Kuehling via Dng wrote:
> Hi,
>
> I upgraded three machines from Devuan ascii to Devuan beowulf. Now one
> of these machines has non-deterministic problems during startup that
> look like being related to eudev. As that machine was originally
> derived by cloning the hard-disk of another machine that works fine, I
> suspect this issue is not a general configuration problem that I caused.

On the contrary, unless the machines hardware are identical in every way
down to the revision level and firmware versions..even then BIOS or UEFI
settings can catch out the unwary. By far the best way is to only transfer
the package cache (/var/cache/apt/archives) and let each system install
them itself.

I suspect this issue may be a general configuration problem that you caused.

Too much mud clouding the waters to see clearly!
Start investigating again after purging *all* un-required packages for
non-existant hardware (lp for a start), then rebuild the initramfs and
*cold* reboot.

You may have half-cloned (exists in config but not in hardware) a UUID!

>
> That machine previously worked flawlessly with Devuan ASCII.
>
> The symptoms are as follows:
>
> This happens around half the times I reboot the machine. In the other
> half of the cases everything works as usual. However, if it fails, this
> happens:
>
> - during startup, the LVM init.d-script hangs for a very long time
> repeatedly outputting errors of the form
>
> WARNING: Device /dev/xxx not initialized in udev database even after
> waiting 10000000 microseconds.
>
> - none of the alsa sound driver modules is loaded during boot, so I
> end up with a Desktop that does not support audio
>
> - the amdgpu module is not loaded, so not 3-D acceleration
>
> Comparing dmesg output between successful startup and failed startup,
> the following markedly differs:
>
> In both cases, the line 'udevd[558]: starting eudev-3.2.7' appears twice
> in the dmesg output. Differences happen after the second time it is
> printed. In the "failed boot" case it merely outputs
>
> xxxx udevd[558]: starting eudev-3.2.7
> xxxx lp: driver loaded but no devices found
> xxxx ppdev: user-space parallel port driver
>
> In the succesful case it outputs a whole lot of ACPI-related stuff and
> loads the sound-related modules:
>
> xxxx udevd[560]: starting eudev-3.2.7
> xxxx acpi_cpufreq: overriding BIOS provided _PSD data
> xxxx input: Power Button as /devices/LNXSYSTM:00/LNXSYBUS:00/PNP0C0C:00/input/input6
> xxxx ACPI: Power Button [PWRB]
> xxxx input: Power Button as /devices/LNXSYSTM:00/LNXPWRBN:00/input/input7
> xxxx ACPI: Power Button [PWRF]
> xxxx ACPI: Video Device [GFX0] (multi-head: yes rom: no post: no)
> xxxx acpi device:00: registered as cooling_device4
> xxxx input: Video Bus as /devices/LNXSYSTM:00/LNXSYBUS:00/PNP0A08:00/LNXVIDEO:00/input/input8
> xxxx sp5100_tco: SP5100/SB800 TCO WatchDog Timer Driver
> xxxx sp5100-tco sp5100-tco: Using 0xfeb00000 for watchdog MMIO address
> xxxx sp5100-tco sp5100-tco: initialized. heartbeat=60 sec (nowayout=0)
> [..]
> xxxx input: PC Speaker as /devices/platform/pcspkr/input/input9
> xxxx udevd[626]: failed to execute '/lib/udev/mtp-probe' 'mtp-probe
> /sys/devices/pci0000:00/0000:00:10.0/usb2/2-3 2 3': No such file or
> directory
> [..]
> xxxx ACPI Error: Needed [Buffer/String/Package], found [Integer]
> 000000003089d4ef (20180810/exresop-560)
> [..]
> xxxx snd_hda_intel 0000:00:01.1: Force to non-snoop mode
> xxxx kvm: disabled by bios
> xxxx MCE: In-kernel MCE decoding enabled.
> xxxx input: HDA ATI HDMI HDMI/DP,pcm=3 as /devices/pci0000:00/0000:00:01.1/sound/card0/input11
> [..]
>
> Any ideas what is happening here and how to workaround it? Or any ideas
> how to further diagnose this problem?
>
> Thanks for any help,
>
> cheers,
>
> David
> _______________________________________________
> Dng mailing list
> Dng@???
> https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
>