systemd questions

[Simo Sorce]

On Thu, 2011-05-19 at 02:06 +0100, Matthew Garrett wrote:
> On Wed, May 18, 2011 at 07:42:17PM -0400, Simo Sorce wrote:
> 
> > > I am pretty sure we don't want to run Java programs at late boot, as
> > > root. This would be really bad.
> > 
> > You know, it's not like there is a choice for many models ...
> 
> That's really not a given. For anything short of us having to send http 
> requests, there's no fundamental reason why this can't be done 
> in-kernel. For cases where we do have to send http requests, well, 
> that's kind of a separate issue (you want to run java after you've 
> unmounted all the filesystems?)

While most UPSs just use serial commands over USB that's not a given.
Some models use more or less documented standards, so you *could*
implement something in kernel I guess.

Some other, more data-centered UPSs that handle multiple machines use
completely proprietary protocols over ethernet for example.

It would be nice to be able to have all in kernel neatly and what not,
but that's not going to happen soon.

Meanwhile we need decent hooks so we can keep using existing models.

> > It is not necessary, the hook can be called before you do that last
> > operation, the driver will command the UPS to wait long enough for that
> > operation to finish in time in 99% of the cases.
> > For the remaining 1% of the cases, admins will have to re-tune the delay
> > once power comes up and they find it wasn't enough time after all.
> 
> Or we could work on fixing the problem. We're not talking about an 
> incredibly large maount of code.

We are talking about unknown code shipped with unknown devices that has
been working for decades. I do not know how much it is, but it is
certainly not a trivial amount.

And it is not a problem we can solve in a short time, so pretty please,
provide a decent hook and let's move on, there are better things on
which to reach "perfection" with systemd, this is not one of them, right
now.

Simo.

-- 
Simo Sorce * Red Hat, Inc * New York