Complex Event Config

[Scott James Remnant]

On Thu, 2007-04-26 at 00:52 +0200, Mildred wrote:

> Le Tue 24/04/2007 à 15:53 Scott James Remnant à écrit:
> > Background and the current specification can be found here:
> > 
> > https://blueprints.beta.launchpad.net/upstart/+spec/complex-event-config
> > 
> 
> It's me or you're giving us a link to the beta version of launchpad ?
> Anyway, that's not important.
> 
> > Note that this has been effectively rejected since it does not solve
> > the problems as effectively as we might like.
> 
> Well, I don't really understand that. reading the spec, it seems to
> address issues we currently have with the simple on stanza. I know
> there are unresolved issues but I can't understand these... But if you
> say that it is not good enough, I must agree. I'm using upstart only
> for few days (well, maybe a week).
> 
The concept certainly lets you combine events in interesting ways;
additional operators are easy to add as well, for example you might add
a "then" operator which means "LHS, and if the job doesn't start, every
time the RHS happens until the job starts"

	start on block-device-added then filesystem-mounted

> > This already tells us the important things about how events and jobs
> > interact, that the complex-event-config specification must address.
> > 
> >  - when are new instances spawned?
> >    (currently "when any start event listed is received by the master
> >     job")
> > 
> >  - when is the job goal changed?
> >    (currently "when any event listed is received")
> > 
> >  - when is the event deemed handled?
> >    (currently "when the job reaches a rest state")
> > 
> >  - when is the event information discarded?
> >    (currently "when the event is deemed handled", soon to be "when the
> >     job goal is changed")
> > 
> 
> The current behaviour (or soon to be) seems something good. Except we
> should allow complex combinations of events. 
> 
> Why not just keep the current behaviour and just adding operators
> between events ?
> 
So the problem then is the following:

	start on starting apache and starting postgresql

What happens if PostgreSQL is never started?  The first event would
still be blocking, so Apache would never start either.

(And maybe that's the right thing, since this job is probably why you're
using Apache in the first place, but it needs thinking about).

Scott
-- 
Have you ever, ever felt like this?
Had strange things happen?  Are you going round the twist?
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : https://lists.ubuntu.com/archives/upstart-devel/attachments/20070426/ad812007/attachment.pgp