[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [libvirt] [Xen-devel] [PATCH 00/12] libxl: fork: SIGCHLD flexibility



On Thu, Jan 30, 2014 at 09:14:01AM -0700, Jim Fehlig wrote:
> Daniel P. Berrange wrote:
> > On Mon, Jan 27, 2014 at 06:39:36PM -0700, Jim Fehlig wrote:
> >   
> >> [Adding libvirt list...]
> >>
> >> Ian Jackson wrote:
> >>     
> >>> Jim Fehlig writes ("Re: [Xen-devel] [PATCH 00/12] libxl: fork: SIGCHLD flexibility"):
> >>>       
> >
> > BTW, I meant to ask before - what is the SIGCHLD reference about in the
> > subject line ?
> >   
> 
> This is related to child processes started by libxl.  E.g. running a
> bootloader when creating PV VMs, running a save/restore helper when
> saving/restoring a VM, etc.
> 
> > libvirt drivers that live inside libvirtd should never use or rely on
> > the SIGCHLD signal at all. All VM processes started by libvirtd ought
> > to be fully daemonized so that their parent is pid 1 / init. This ensures
> > that the libvirtd daemon can be restarted without all the VMs getting
> > reaped. Once the VMs are reparented to init, then libvirt or library
> > code it uses has no way of ever receiving SIGCHLD.
> >   
> 
> Nod.  VMs are "daemonized", in the context of Xen.  Once running,
> libvirtd can be restarted without reaping them.

Great, thanks for clarifying, this sounds fine then.

Daniel
-- 
|: http://berrange.com      -o-    http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org              -o-             http://virt-manager.org :|
|: http://autobuild.org       -o-         http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org       -o-       http://live.gnome.org/gtk-vnc :|


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]