Re: New libgtop maintainer

[Stefano Facchini <stefano facchini gmail com>]

Il giorno dom, 11/08/2013 alle 14.20 +0200, Maciej Piechotka ha scritto:
> On Sun, 2013-08-11 at 14:03 +0200, Stefano Facchini wrote:
> > Il giorno dom, 11/08/2013 alle 06.45 -0400, Jasper St. Pierre ha
> > scritto:
> >
> > > So, I wonder if it makes sense to stop generating libgtop and instead
> > > just focus on a solid, easily understood codebase. I never really
> > > understood why we had a client/daemon split, either; it doesn't seem
> > > that we're doing anything too fancy on either side. Is it that we
> > > require root for reading some of the files? Should we move to a system
> > > DBus service instead?
> >
> > I think that root access is required if we want to implement monitoring
> > of:
> >   * per process disk activity (à la iotop)
> >   * per process network usage (à la nethogs)
> >
> > That said, a DBus service should be perfectly fine for these features.
>
> As a question - what about timeouts? Usually gnome-system-monitor is
> useful when there is heavy I/O, CPU usage or swapping and in such cases
> dbus timeouts can and do happen.
Yes you're right, when there's high disk activity dbus can be slow. I
was more thinking of network usage, because it's the only one I started
prototyping as DBus service :)

> Moving into more complex area - since application handling on Linux
> moves to systemd+cgroups would it make sense to get information per
> cgroup rather then per-process (+ nice user-readable name such as "Web"
> or "Epiphany" rather then "/usr/libexec/WebKitPluginProcess")? Possibly
> something less readable for systems without cgroup-like session
> handling.