[systemd-devel] systemd .device units timing out

[Avi Kivity]

On 06/15/2011 12:17 PM, Avi Kivity wrote:
> On 06/14/2011 06:17 PM, Lennart Poettering wrote:
>> On Mon, 06.06.11 16:01, Avi Kivity (avi at redhat.com) wrote:
>>
>> >  I have a system which was upgraded to F15 and thus systemd.  During
>> >  the boot process it times out starting various .device units, like
>> >  sda1.device.  If I let it alone it will boot in about a week or so,
>> >  but meanwhile it is unusable.  I am able to start it up using
>> >  emergency.target and various hacks, but it won't boot cleanly.
>> >
>> >  Any hints?
>>
>> Hmm, systemd only sees devices properly marked in udev with the systemd
>> tag. It would be interesting to find out if the block devices are
>> properly marked in the udev db.
>>
>> Can you paste the output of "udevadm info -qall -p
>> /sys/class/block/sda1" after this happened? This should tell us what
>> data might be missing in the db.
>>
>
> It's hard to reproduce, since I can't stop device units:

My machine died so I got to reproduce this.  Indeed, booting into 
emergency.target, the systemd tag is missing.  Restarting udevd got it 
tagged correctly, and things work.

(an oddity - systemctl start dbus.service gets the gdm greeter running.  
Why?)

-- 
error compiling committee.c: too many arguments to function