Re: systemd - excessive session-creation time

[Zenaan Harkness <zen@freedbms.net>]

On 5/18/14, Sven Joachim <svenjoac@gmx.de> wrote:
> On 2014-05-18 07:09 +0200, Andrei POPESCU wrote:
>> On Du, 18 mai 14, 12:49:08, Zenaan Harkness wrote:
>>> Although initial tty session (from boot) is quicker than with
>>> sysvinit, additional tty sessions are very slow to start, in the order
>>> of 5 seconds (which seems like an eternity).
>> ...
>>> Any ideas why this might be happening?
>>
>> As far as I recall consoles are spawned only as needed (when you
>> actually switch to that VT)
>
> Nitpick: the consoles are always there, it's the program which is
> running on them (getty) which is spawned on demand.

Is the number of consoles fixed by the kernel (eg 6, 7 or 8)?

>> and I think there is an option to always
>> enable them. Hope this puts you on the right track.
>
> I suspect a problem with hostname resolution (with the default
> /etc/issue getty needs to know the hostname).

The delay would align with hostname resolution problems, but I have
configured a hostname, and 'hostname' runs instantly:

$ hostname
x220a02

$ grep x220a02 /etc/hosts
127.0.1.1      x220a02

May be the hostname must resolve to 127.0.0.1?

Or could there be some other hostname problem?

Again, the key problem I am facing at this point, is merely an
excessive delay on new getty spawns, and a similar delay on loggin out
of X (delay to get back to a getty login prompt), and a similar delay
on shutdown - there appears to be this consistent delay in the
transition from X -> vt, which is unrelated to getty though, as I see
it, since this same problem occurs when I try to shutdown/halt,
mid-shutdown.

Again, we're talking startx scenario.

TIA
Zenaan