Laurent Bercot <ska-devel@???> writes:
> On 25/09/2015 17:29, Simon Hobson wrote:

[...]

>> Of course, regardless of what system or definitions you use - if a
>> service then dies then you have a problem. IMO, "it might die at some
>> indeterminate time" isn't an excuse for not trying to get the "start
>> stuff up" part right.
>
> Apparently Rainer disagrees with that, and seems to think that since
> you can't get a 100% reliable system, it's useless to get the common
> case working as smoothly as possible. I've stopped trying to convince
> him.

Since this is not some technical statement but one about me, I think I
have to address it somehow: We weren't discussing whether or not "get
the common case working as smoothly as possible" makes sense[*] but I
asked for examples of real 'server dependencies' where performing a
topological sort in order to work out a (partial) order for 'starting
servers' would be technically useful. In my opinion, this cannot be
because 'starting the server at time X' is a necessary but not a sufficient
precondition for 'server will serve requests at time Y, Y > X'. The
situation is also going to sort itself out quickly, that is, within
seconds and a system will spend most of its lifetime with being in some
sort of steady state performing whatever its function happens to be, not
with "being started".

[*] In this generality, such a discussion seems useless to me: Eg, the
    common case for a human being is that it can read and write Chinese,
    hence, use of any other system of writing ought to be avoided.