feature/geo-rep: Fix to retain pause state of gsyncd on restart

On soft reboot, geo-rep monitor is writing 'faulty' into status
file. It should not do it if previous state is paused as
glusterd depend on the state file on node restart.

Change-Id: Idd45abf13350b087371935f1b4f6e1a346433d27
BUG: 1101410
Signed-off-by: Kotresh H R <khiremat@redhat.com>
Reviewed-on: http://review.gluster.org/8097
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Venky Shankar <vshankar@redhat.com>
Tested-by: Venky Shankar <vshankar@redhat.com>

2 files changed, 18 insertions, 14 deletions

diff --git a/geo-replication/syncdaemon/monitor.py b/geo-replication/syncdaemon/monitor.py
index 0bde216d761..5f4745501ec 100644
--- a/geo-replication/syncdaemon/monitor.py
+++ b/geo-replication/syncdaemon/monitor.py
@@ -73,11 +73,10 @@ class Monitor(object):
     """class which spawns and manages gsyncd workers"""
 
     ST_INIT = 'Initializing...'
-    ST_INIT_PAUSE = 'Initializing...(Paused)'
     ST_STABLE = 'Stable'
     ST_FAULTY = 'faulty'
     ST_INCON = 'inconsistent'
-    _ST_ORD = [ST_STABLE, ST_INIT, ST_INIT_PAUSE, ST_FAULTY, ST_INCON]
+    _ST_ORD = [ST_STABLE, ST_INIT, ST_FAULTY, ST_INCON]
 
     def __init__(self):
         self.lock = Lock()
@@ -97,8 +96,17 @@ class Monitor(object):
             if state != old_state:
                 self.set_state(state)
         else:
-            logging.info('new state: %s' % state)
             if getattr(gconf, 'state_file', None):
+                # If previous state is paused, suffix the
+                # new state with '(Paused)'
+                try:
+                    with open(gconf.state_file, "r") as f:
+                        content = f.read()
+                        if "paused" in content.lower():
+                            state = state + '(Paused)'
+                except IOError:
+                    pass
+                logging.info('new state: %s' % state)
                 update_file(gconf.state_file, lambda f: f.write(state + '\n'))
 
     @staticmethod
@@ -129,10 +137,7 @@ class Monitor(object):
         due to the keep-alive thread)
         """
 
-        if gconf.pause_on_start:
-            self.set_state(self.ST_INIT_PAUSE, w)
-        else:
-            self.set_state(self.ST_INIT, w)
+        self.set_state(self.ST_INIT, w)
 
         ret = 0
 
@@ -310,5 +315,11 @@ def distribute(*resources):
 
 
 def monitor(*resources):
+    # Check if gsyncd restarted in pause state. If
+    # yes, send SIGSTOP to negative of monitor pid
+    # to go back to pause state.
+    if gconf.pause_on_start:
+        os.kill(-os.getpid(), signal.SIGSTOP)
+
     """oh yeah, actually Monitor is used as singleton, too"""
     return Monitor().multiplex(*distribute(*resources))
diff --git a/geo-replication/syncdaemon/resource.py b/geo-replication/syncdaemon/resource.py
index eb6c87b102b..b537ff65003 100644
--- a/geo-replication/syncdaemon/resource.py
+++ b/geo-replication/syncdaemon/resource.py
@@ -1294,13 +1294,6 @@ class GLUSTER(AbstractUrl, SlaveLocal, SlaveRemote):
                 logging.debug("Changelog register failed: %s - %s" %
                               (e.errno, e.strerror))
 
-            # Check if gsyncd restarted in pause state. If
-            # yes, send SIGSTOP to negative of monitor pid
-            # to go back to pause state.
-            if gconf.pause_on_start:
-                os.kill(-os.getppid(), signal.SIGSTOP)
-                gconf.pause_on_start = False
-
             # oneshot: Try to use changelog history api, if not
             # available switch to FS crawl
             # Note: if config.change_detector is xsync then