[pve-devel] applied: [PATCH manager v2] Fix #1386: track postfix daemon status via postfix at -

[Wolfgang Bumiller]

applied

On Thu, May 18, 2017 at 11:10:52AM +0200, Emmanuel Kasper wrote:
> ---
> IMHO #1387 has a different scope. This patch here deals with 
> fixing the default reporting when a the status of one of our
> default daemon is reported in a different unit.
> #1387 is about *manually* changing the default startup of a 
> service to socket activation, and then not getting the proper
> status afterwards in the status api call for this service.
> 
>  PVE/API2/Services.pm | 9 ++++++++-
>  1 file changed, 8 insertions(+), 1 deletion(-)
> 
> diff --git a/PVE/API2/Services.pm b/PVE/API2/Services.pm
> index 3ddf5f56..d031a3a6 100644
> --- a/PVE/API2/Services.pm
> +++ b/PVE/API2/Services.pm
> @@ -35,9 +35,16 @@ my $service_name_list = [
>      'systemd-timesyncd',
>      ];
>  
> +# since postfix package 3.1.0-3.1 the postfix unit is only here to
> +# manage subinstances, of which the  default is called "-".
> +# This is where we look for the daemon status
> +my $unit_extra_names = {
> +    postfix => 'postfix at -'
> +};
> +
>  my $get_full_service_state = sub {
>      my ($service) = @_;
> -
> +    $service = $unit_extra_names->{$service} if $unit_extra_names->{$service};
>      my $res;
>      
>      my $parser = sub {
> -- 
> 2.11.0