:: Re: [DNG] Detailed technical treati…
Góra strony
Delete this message
Reply to this message
Autor: Didier Kryn
Data:  
Dla: dng
Temat: Re: [DNG] Detailed technical treatise of systemd
Le 03/11/2015 17:24, Rainer Weikusat a écrit :
> Didier Kryn <kryn@???> writes:
>>
>>      I agree with you, and it was the first point in my mail, that the
>> servers should be able to cope with outages.
> That's not a matter of "should": They have to. Even if it's believed
> they're just using local IPC[*].


     Yes, but let's consider that, maybe, some do not.


>
>> However let's not be extremists.
> Feel free to consider yourself 'an extremist', ie, someone fighting
> (often violently) for a certain policy regardless of reason but please
> don't label me in this way: The issue is technical and not political and
> the proposed solution doesn't work.


     Extremists always have reasons - excellent, according to them. I 
sometimes read opinions which I consider as technically extremist. I 
would rank in this category the will to impose one single init system, 
wether it be systemd or any other, or the will to impose one single text 
editor.


     If you have taken the word against you, please consider I withdraw 
it, because it wasn't my intention. I was just meaning to express my 
favour towards a balanced policy. Let me explicite it:


     - Encourage authors to make outage-aware servers, which can then be 
started in parallel;
     - Provide a supervisor able to handle dependencies for the 
non-outage-aware, with a trivial readyness notification method.


>
>> In most cases this is going to work, because
>> Time_of_check_to_time_of_use issues do not arise all the time.
> That's a property of any problem caused by a race condition: The time
> when the latent problem will manifest itself can't (easily) be
> predicted.
>
>


     This is why people have been able to live with sysv-init for many 
decades :-)



Le 03/11/2015 20:49, Steve Litt a écrit :
> There's an init system called s6-rc that just came out or is just
> coming out. From what I understand, it starts stuff sequentially or in
> parallel, with very sophisticated process dependency checking
> available, and I think it has an extremely simple way for processes to
> communicate readiness.


     That's exactly what I have in mind :-) And I think it can become 
the new reference for start up.


     Didier