-Fix problem where extremely large Bladecenter configurations could fail to execute a hardware management command

git-svn-id: https://svn.code.sf.net/p/xcat/code/xcat-core/trunk@1912 8638fb3e-16cb-4fca-ae20-7b5d299a9bcd

xCAT-server/lib/xcat/plugins/blade.pm

@@ -1613,7 +1613,7 @@ sub process_request {
   }
   my $sub_fds = new IO::Select;
   foreach $mpa (sort (keys %mpahash)) {
-    while ($children > $blademaxp) { sleep (0.1); }
+    while ($children > $blademaxp) { forward_data($callback,$sub_fds); }
     $children++;
     my $cfd;
     my $pfd;