Re: [Firebird-devel] official install script on ubuntu 10.10 does have some errors , must be corrected (I will add a bug later today)

[Alex P. <pes...@ma...>]

 On 11/03/10 13:41, marius adrian popa wrote:
> On Wed, Nov 3, 2010 at 10:16 AM, Alex Peshkoff <pes...@ma...> wrote:
>>  On 11/03/10 02:32, Philippe Makowski wrote:
>>> Alex Peshkoff  [2010-10-29 16.24] :
>>>>> Also in we have to replace System V style init scripts with upstart
>>>>> jobs in Ubuntu ...
>>>> Hops, one more format of daemon startup. Sometimes I start to think
>>>> better about windows with it's well documented services.
>>> OpenSuse and Fedora are moving to Systemd
>>> http://www.freedesktop.org/wiki/Software/systemd
>>> http://0pointer.de/blog/projects/systemd.html
>> This is much more serious change - this is not another format of startup
>> scripts, but replacement of init (pid 1). Looks very interesting, but
>> unfortunately is still in experimental state. And requires reworking of
>> not only startup script (appears it's not needed at all), but daemon
>> code too, like getting socket to listen from environment instead opening
>> it ourself. Quite possible that sooner or later we will not need other
>> startup types except systemd.
> seems that it will be in fedora 15  by default
> http://fedoraproject.org/wiki/Releases/15/FeatureList
> with the fail over to upstart if it will not be so stable and removed
> from features lists

gentoo currently marks this project as highly experimental

> I wish they have made an init systems based on python scripts and
> replace the actual bashisms

May be I missed something, but looks like there will not be startup
scripts at all. Configuration of required services is known to systemd,
and it starts them when decides it's needed. BTW, something (more or
less) similiar does xinetd.

> These new init systems are mostly for Desktop systems (booting < 15s)

Making boot a few times faster is useful for any kind of systems. What's
bad if notebook starts faster? As a minimum it saves power, giving a
user one additional minute to work. Certainly, servers are not restarted
so often, but if after long power outage data farm can be started one
minute faster that's also good.