watchquagga: Signal when we are actually up and running

When Quagga is starting up, it is returning immediately.
This is leaving us in a state where systemd believes
Quagga is up and running, while the sytem might actually
not have restarted the code yet.

Modify the code so that when watchquagga starts up
it doesn't start communicating with systemd until
such time that it detects that all daemons are
running.

Additionally modify watchquagga to touch a
file in /var/run/quagga/ that the /usr/lib/quagga/quagga
script looks for for 10 seconds.  If it finds this
Quagga started file then we know watchquagga
has successfully communicated with all daemons.
If after 10 seconds we haven't communicated
with Quagga, continue on for the start and let the
normal start failure code work.

Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>

2 files changed, 35 insertions, 3 deletions

diff --git a/tools/quagga b/tools/quagga
index 83dfb6302..e8595d787 100755
--- a/tools/quagga
+++ b/tools/quagga
@@ -33,7 +33,6 @@ else
     SSD=`which start-stop-daemon`
 fi
 
-echo ${SSD}
 # Print the name of the pidfile.
 pidfile()
 {
@@ -114,12 +113,23 @@ start()
                echo -n " $1"
            fi
 
+           if [ -e /var/run/quagga/watchquagga.started ] ; then
+             rm /var/run/quagga/watchquagga.started
+           fi
             ${SSD} \
                 --start \
                 --pidfile=`pidfile $1` \
                 --exec "$D_PATH/$1" \
                 -- \
                 "${watchquagga_options[@]}"
+            for i in `seq 1 10`;
+            do
+              if [ -e /var/run/quagga/watchquagga.started ] ; then
+		  break
+              else
+                  sleep 1
+              fi
+            done
         elif [ -n "$2" ]; then
           echo -n " $1-$2"
            if ! check_daemon $1 $2 ; then
@@ -502,8 +512,8 @@ case "$1" in
         if [ "$2" != "watchquagga" ]; then
           start_prio 10 $dmn
         fi
-        vtysh_b
         start_watchquagga
+        vtysh_b
         ;;
 
     1|2|3|4|5|6|7|8|9|10)
diff --git a/watchquagga/watchquagga.c b/watchquagga/watchquagga.c
index e882653e3..cb9d50ff8 100644
--- a/watchquagga/watchquagga.c
+++ b/watchquagga/watchquagga.c
@@ -682,6 +682,28 @@ handle_read(struct thread *t_read)
   return 0;
 }
 
+/*
+ * Wait till we notice that all daemons are ready before
+ * we send we are ready to systemd
+ */
+static void
+daemon_send_ready (void)
+{
+  static int sent = 0;
+  if (!sent && gs.numdown == 0)
+    {
+#if defined (HAVE_CUMULUS)
+      FILE *fp;
+
+      fp = fopen("/var/run/quagga/watchquagga.started", "w");
+      fclose(fp);
+#endif
+      zlog_notice ("Watchquagga: Notifying Systemd we are up and running");
+      systemd_send_started(master, 0);
+      sent = 1;
+    }
+}
+
 static void
 daemon_up(struct daemon *dmn, const char *why)
 {
@@ -689,6 +711,7 @@ daemon_up(struct daemon *dmn, const char *why)
   gs.numdown--;
   dmn->connect_tries = 0;
   zlog_notice("%s state -> up : %s",dmn->name,why);
+  daemon_send_ready();
   if (gs.do_ping)
     SET_WAKEUP_ECHO(dmn);
   phase_check();
@@ -1284,7 +1307,6 @@ main(int argc, char **argv)
       
   gs.restart.interval = gs.min_restart_interval;
   master = thread_master_create();
-  systemd_send_started (master, 0);
   signal_init (master, array_size(my_signals), my_signals);
   srandom(time(NULL));