NethServer 6

Symptom
Ejabberd creates a crash dump every time the Dashboard page is loaded.

Step to reproduce
Reload the Dashboard page, check a new crash file (erl_crash_<time>.dump) is created under /var/log/ejabberd directory.

Analysis
The Dashboard checks service running status on each load using /usr/libexec/nethserver/read-service-status script.
This script relies on NethServer::Service library which implements a call to the underling system (init or upstart).
The _silent_system handle this check trying to avoid output on stderr and stdout, but current implementation seems to have problems with init scripts spawning multiple children.

Very simple proposed patch (tested):

diff --git a/lib/perl/NethServer/Service.pm b/lib/perl/NethServer/Service.pm
index 6571d43..1e645e7 100644
--- a/lib/perl/NethServer/Service.pm
+++ b/lib/perl/NethServer/Service.pm
@@ -397,14 +397,7 @@ sub _get_command
 #
 sub _silent_system
 {
-    if(fork()) {
-       wait();
-       return $?;
-    } else {
-       open(STDOUT, '/dev/null');
-       open(STDERR, '/dev/null');
-       exec(@_);
-    }
+    return system("@_ &>/dev/null")
 }

 #