libvirt-cim-bounces@redhat.com wrote on 2008-09-11 04:46:16:

> >  >
> >  > I gave them some more thought.. adding an arbitrary sleep in a test
> >  > isn't really a good idea - because systems run at varying speeds, so you
> >  > can't guarantee how long you'll need to sleep.
> >  >
> >  > A better approach would be to poll to see if there are no guests... if
> >  > there are no guests defined, then run the test, otherwise keep polling.
> >  >   You could poll for 10 seconds or so.  So something like:
> >  >
> >  > timer_count = 10
> >  >
> >  > if count < timer_count:
> >  >    #Get list of defined/active guests on the system
> >  >    #If no guests are defined
> >  >        break
> >  >    sleep(1)
> >  >
> >  > Does this sound reasonable?
> >
> >   Agree with you.
> >   If there is still define guest on the system after waiting for
> >   10 seconds, then shall we have to make the timer_count longer?
> >
>
> The problem here is that it's possible the user has other defined guests
> in the system.  So we can't keep polling for there to be no guests.  The
> best we can do is poll for a given amount of time.  If the system still
> has guests on it after the timer is up, then we need to skip the test.
>
  Done. Patch on the way.

> --
> Kaitlin Rupert
> IBM Linux Technology Center
> kaitlin@linux.vnet.ibm.com
>
> _______________________________________________
> Libvirt-cim mailing list
> Libvirt-cim@redhat.com
> https://www.redhat.com/mailman/listinfo/libvirt-cim