/var/log/messages

[Ralf Corsepius]

On 12/22/2011 10:06 AM, Reindl Harald wrote:
>
>
> On 22.12.2011 07:27, Joe Zeff wrote:
>> On 12/21/2011 11:10 PM, Ralf Corsepius wrote:
>>> I would call it a functional regression having been introduce during by
>>> the sysvinit/upstart->systemd transition.
>>>
>>> If it has not been BZ'd before, it should be BZ'ed.
>>
>> I wouldn't mind being the reporter, but I'd like to know if anybody else sees
>> this or not.  I'll try to find time to fire up my laptop tomorrow and check there
>> as well, but I'd also appreciate it if the rest of you would check and let me know
>> if it's been off.
>
> this si all correct
>
> but normally you should make a list which services are running
> before the upgrade and verify that there are none missing after
> the upgrade or services you did disable before were silently
> enabled again
I would expect an upgrade to preserve those service which were running 
before the upgrade. This is what upgrading to systemd was supposed to 
do. The fact it doesn't therefore qualifies as a bug.

> what do you do if some bug disables sshd while you make the
> upgrade over ssh? after a blindly reboot the amchine is away
> for you!

... why limit this to ssh? There are many possibilities for this to 
happen. AFAICT, systemd is much more sensitive to such issues than 
sysvint/upstart used to be.

Ralf