[libvirt] [libvirt-tck 2/3] Add module for hooks testing

[Osier]

----- "Daniel P. Berrange" <berrange at redhat.com> wrote:

> On Mon, Oct 18, 2010 at 07:18:08AM +0800, Osier Yang wrote:
> > To test daemon, qemu, lxc hook.
> > 
> > * lib/Sys/Virt/TCK/Hooks.pm
> > ---
> >  lib/Sys/Virt/TCK/Hooks.pm |  262
> +++++++++++++++++++++++++++++++++++++++++++++
> >  1 files changed, 262 insertions(+), 0 deletions(-)
> >  create mode 100644 lib/Sys/Virt/TCK/Hooks.pm
> > 
> > diff --git a/lib/Sys/Virt/TCK/Hooks.pm b/lib/Sys/Virt/TCK/Hooks.pm
> 
> > +sub libvirtd_status {
> > +    my $self = shift;
> > +    my $status = `service libvirtd status`;
> > +    my $_ = $status;
> > +
> > +    if (/running/) {
> > +        $self->{libvirtd_status} = 'running';
> > +    } elsif (/stopped/) {
> > +        $self->{libvirtd_status} = 'stopped';
> > +    }
> > +
> > +    return $self;
> > +}
> 
> > +sub service_libvirtd {
> > +    my $self = shift;
> > +    my $action = $self->{action};
> > +
> > +    truncate $self->{log_name}, 0 if -f $self->{log_name};
> > +
> > +    die "failed on $action daemon" if system "service libvirtd
> $action"; 
> > +
> > +    $self->libvirtd_status;
> > +}
> 
> Is there any way we can avoid having to start/stop libvirtd
> for this testing ?  The general goal of the TCK is that it
> is testing an existing deployment, so it should be expecting
> that libvirtd is already up & running in a desired configuration.
> 
> If we have to stop/start libvirtd, then the test script using
> these APIs will need to be protected to make sure it only
> runs when used with 'qemu:///system' or 'lxc://'.  ie is skipped
> with qemu:///session or vmware, or virtualbox, etc
> 

For daemon hook testing, It's neccessary to start/stop/restart the 
libvirtd. Otherwise we can't see if the hook script is invoked or not.
It doesn't relate to which hypervisor driver is used..

> > +
> > +sub compare_log {
> > +    my $self = shift;
> > +    
> > +    my $expect_log = $self->{expect_log};
> > +    my $log_name = $self->{log_name};
> > +
> > +    open LOG, "< $log_name" or die "failed on opening $log_name:
> $!";
> > +    
> > +    my @lines = <LOG>;
> > +
> > +    return 0 unless @lines;
> > +
> > +    chomp foreach @lines;
> > +    my $actual_log = join "\n", @lines;
> > +
> > +    close LOG;
> 
> Little perl black magic tip for you.... 
> 
> If you want to read the entire file contents into a single
> string, then you can do
> 
>   open LOG, "<$log_name";
>   local $/ = undef;
>   my $actual_log = <LOG>;
>   close LOG;
> 
> '$/' is the line separator. By setting it to 'undef' we tell
> Perl that there is no line separator, so it will immediately
> read until end of file :-)  BTW see 'man perlvar' for this 
> particular example
> 

cool trick.. will update it.. thanks.. :-)

> 
> Daniel
> -- 
> |: Red Hat, Engineering, London    -o-  
> http://people.redhat.com/berrange/ :|
> |: http://libvirt.org -o- http://virt-manager.org -o-
> http://deltacloud.org :|
> |: http://autobuild.org        -o-        
> http://search.cpan.org/~danberr/ :|
> |: GnuPG: 7D3B9505  -o-   F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B
> 9505 :|