diff --git a/Netdisco/bin/netdisco-daemon b/Netdisco/bin/netdisco-daemon
index c6986057..e8112538 100755
--- a/Netdisco/bin/netdisco-daemon
+++ b/Netdisco/bin/netdisco-daemon
@@ -29,14 +29,24 @@ sub gd_preconfig {
   # used for locking jobs in central Pg queue
   $gd->{nd_host} = hostfqdn;
 
-  # deploy the daemon's local DB schema
-  $gd->deploy_daemon_db;
+  set(daemon_pollers => 2)
+    if !defined setting('daemon_pollers');
+  set(daemon_interactives => 2)
+    if !defined setting('daemon_interactives');
 
-  # sync with netdisco Pg DB
-  $gd->netdisco_db_sync;
-
-  # init Parallel::Prefork
-  $gd->bootstrap_prefork;
+  # need to do this after setting defaults
+  $pp = Parallel::Prefork->new(
+    max_workers => (1 + setting('daemon_pollers')
+                      + setting('daemon_interactives')),
+    spawn_interval => 2,
+    before_fork => \&set_next_worker_role,
+    after_fork  => \&register_worker,
+    on_child_reap => \&unregister_worker,
+    trap_signals => {
+      TERM => 'TERM',
+      INT  => 'TERM',
+    },
+  );
 
   # do not remove this line - required by Daemon::Generic
   return ();
@@ -50,9 +60,11 @@ sub gd_run_body {
     if $pp->signal_received =~ m/^(?:TERM|INT)$/;
 
   $pp->start(sub {
+    print STDERR ">>> new $next_role worker starting...\n";
     with "Netdisco::Daemon::Worker::$next_role";
-    print STDERR ">>> new $next_role worker started.\n";
+    $gd->worker_begin if $gd->can('worker_begin');
     $gd->worker_body;
+    $gd->worker_end if $gd->can('worker_end');
   });
 
   # I don't think Parallel::Prefork ever returns from start()
@@ -116,63 +128,6 @@ END {
   $pp->wait_all_children;
 }
 
-sub deploy_daemon_db {
-  my $gd = shift;
-  my $daemon = schema('daemon');
-
-  try {
-      $daemon->storage->dbh_do(sub {
-        my ($storage, $dbh) = @_;
-        $dbh->selectrow_arrayref("SELECT * FROM admin WHERE 0 = 1");
-      });
-  }
-  catch { $daemon->deploy };
-
-  $daemon->storage->disconnect;
-  if ($daemon->get_db_version < $daemon->schema_version) {
-      $daemon->upgrade;
-  }
-}
-
-sub netdisco_db_sync {
-  my $gd = shift;
-  my $netdisco = schema('netdisco');
-  my $daemon = schema('daemon');
-
-  # on start, any jobs previously grabbed by a daemon on this host
-  # will be reset to "queued", which is the simplest way to restart them.
-
-  my $rs = $netdisco->resultset('Admin')->search({
-    status => "running-$gd->{nd_host}"
-  });
-
-  if ($rs->count > 0) {
-      $daemon->resultset('Admin')->delete;
-      $rs->update({status => 'queued', started => undef});
-  }
-}
-
-sub bootstrap_prefork {
-  my $gd = shift;
-
-  # set defaults
-  set(daemon_pollers => 2) if !defined setting('daemon_pollers');
-  set(daemon_interactives => 2) if !defined setting('daemon_interactives');
-
-  # need to do this after setting defaults
-  $pp = Parallel::Prefork->new(
-    max_workers => (1 + setting('daemon_pollers') + setting('daemon_interactives')),
-    spawn_interval => 2,
-    before_fork => \&set_next_worker_role,
-    after_fork  => \&register_worker,
-    on_child_reap => \&unregister_worker,
-    trap_signals => {
-      TERM => 'TERM',
-      INT  => 'TERM',
-    },
-  );
-}
-
 # nullify this to permit Parallel::Prefork to register handlers instead
 sub gd_setup_signals {}
 
diff --git a/Netdisco/lib/Netdisco/Daemon/Worker/Manager.pm b/Netdisco/lib/Netdisco/Daemon/Worker/Manager.pm
index d0522b28..12bcbf07 100644
--- a/Netdisco/lib/Netdisco/Daemon/Worker/Manager.pm
+++ b/Netdisco/lib/Netdisco/Daemon/Worker/Manager.pm
@@ -9,6 +9,37 @@ use Try::Tiny;
 use Role::Tiny;
 use namespace::clean;
 
+sub worker_begin {
+  my $self = shift;
+  my $daemon = schema('daemon');
+
+  # deploy local db if not already done
+  try {
+      $daemon->storage->dbh_do(sub {
+        my ($storage, $dbh) = @_;
+        $dbh->selectrow_arrayref("SELECT * FROM admin WHERE 0 = 1");
+      });
+  }
+  catch { $daemon->deploy };
+
+  $daemon->storage->disconnect;
+  if ($daemon->get_db_version < $daemon->schema_version) {
+      $daemon->upgrade;
+  }
+
+  # on start, any jobs previously grabbed by a daemon on this host
+  # will be reset to "queued", which is the simplest way to restart them.
+
+  my $rs = schema('netdisco')->resultset('Admin')->search({
+    status => "running-$self->{nd_host}"
+  });
+
+  if ($rs->count > 0) {
+      $daemon->resultset('Admin')->delete;
+      $rs->update({status => 'queued', started => undef});
+  }
+}
+
 sub worker_body {
   my $self = shift;