[hostmonitor] Add pacemaker_node_type option

When running in a container, it might not be possible to use systemd
to verify the status of Corosync and Pacemaker.
In such case, allow the user to choose the stack being used.

Change-Id: I44ce3be6b6fda3834f6df63861b0dcf546da46a1
Co-Authored-By: Radosław Piliszek <radoslaw.piliszek@gmail.com>

masakarimonitors/conf/host.py

@@ -79,6 +79,17 @@ like '5405,5406'.
 The number of port numbers you specify must be equal to the number of
 corosync_multicast_interfaces values and must be in correct order with
 relevant interfaces in corosync_multicast_interfaces.
+'''),
+    cfg.StrOpt('pacemaker_node_type',
+               default='autodetect',
+               choices=('autodetect', 'cluster', 'remote'),
+               help='''
+Using this option, one can avoid systemd checks that would establish whether
+this hostmonitor is running alongside Corosync and Pacemaker (the ``cluster``
+stack) or Pacemaker Remote (the ``remote`` stack).
+
+The default (``autodetect``) ensures backward compatibility and means systemd
+is used to check the stack.
 '''),
 ]
 

masakarimonitors/hostmonitor/host_handler/handle_host.py

@@ -79,11 +79,19 @@ class HandleHost(driver.DriverBase):
         :returns: 0 if normal, 1 if abnormal, 2 if configuration file is
             wrong or neither pacemaker nor pacemaker-remote is running.
         """
-        # Check whether the pacemaker services is normal.
+        if CONF.host.pacemaker_node_type == 'autodetect':
-        corosync_status = self._check_pacemaker_services('corosync')
+            # Check whether the Corosync/Pacemaker services are running.
-        pacemaker_status = self._check_pacemaker_services('pacemaker')
+            corosync_status = self._check_pacemaker_services('corosync')
-        pacemaker_remote_status = self._check_pacemaker_services(
+            pacemaker_status = self._check_pacemaker_services('pacemaker')
-            'pacemaker_remote')
+            pacemaker_remote_status = self._check_pacemaker_services(
+                'pacemaker_remote')
+        else:
+            corosync_status = (
+                CONF.host.pacemaker_node_type == 'cluster')
+            pacemaker_status = (
+                CONF.host.pacemaker_node_type == 'cluster')
+            pacemaker_remote_status = (
+                CONF.host.pacemaker_node_type == 'remote')
 
         if corosync_status is False or pacemaker_status is False:
             if pacemaker_remote_status is False:
@@ -402,11 +410,16 @@ class HandleHost(driver.DriverBase):
                     eventlet.greenthread.sleep(CONF.host.monitoring_interval)
                     continue
 
+                if CONF.host.pacemaker_node_type == 'autodetect':
+                    pacemaker_remote_status = self._check_pacemaker_services(
+                        'pacemaker_remote')
+                else:
+                    pacemaker_remote_status = (
+                        CONF.host.pacemaker_node_type == 'remote')
+
                 # Check the host status is stable or unstable by crmadmin.
                 # It only checks when this process runs on the full cluster
                 # stack of corosync.
-                pacemaker_remote_status = self._check_pacemaker_services(
-                    'pacemaker_remote')
                 if pacemaker_remote_status is False:
                     if self._check_host_status_by_crmadmin() != 0:
                         LOG.warning("hostmonitor skips monitoring hosts.")

releasenotes/notes/hostmonitor-systemd-89696f96a654a918.yaml

@@ -0,0 +1,6 @@
+---
+features:
+  - |
+    Adds ``pacemaker_node_type`` option to ``hostmonitor`` to allow
+    skipping systemd service status checks which are often impossible in
+    containerised deployments.