tito <farmatito@???> writes:
> On Wed, 04 Jun 2025 12:26:19 +0100
> Rainer Weikusat via Dng <dng@???> wrote:
>
>> Rainer Weikusat via Dng <dng@???> writes:
>> > Not very amusing story someone might nevertheless find quite funny:
>> >
>> > I used to run RAID1 system with two RAID devices, one for 'system' and
>> > one for 'data', with half of the array residing on an internal NVME
>> > device and the other on an USB disk. I recently bought a webcam. When I
>> > initially plugged that into an USB port, nothing happened. I then tried
>> > a reboot. This caused the camera to appear (no idea if it also works)
>> > but additionally, convinced something in the initramfs to remove all the
>> > partitions on the USB disk from the corresponding RAID devices, thus
>> > leaving me with two degraded arrays.
>> >
>> > I'll now try to figure out how I can get my missing partitions back ...
>>
>> Simply re-adding the lost partitions to the RAIDs they belonged to via
>>
>> mdadm --manage /dev/mdX --add /dev/sdaX
>>
>> turned out ot work. I now again have to fully functioning RAID devices
>> which get autostarted on boot. Nevertheless, this was an 'unexpected
>> feature' I could easily have done without.
>
> Hi,
> probably the disk was not fast enough to spin up for mdadm
> due to the added cam and/or underpowered during the spin up.
> Solution could be to power it separately with a PSU otherwise
> this will randomly happen again at boot (typical of NAS with many disks
> and underpowered or dying PSU showing disks kicked out at boot).

[...]

> Hope this helps,

It's at least an interesting bit of information, thanks. But considering
that the disk was working by the time I logged in, I consider this a
bug, probably, a timeout that's too aggressive. Should this happen
again, I'll seek to fix that.