Add a new cluster parameter maintain_node_health

This will be used to conditionally enable the watcher node maintenance
feature.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>

lib/bootstrap.py

@@ -197,7 +197,8 @@ def InitCluster(cluster_name, mac_prefix,
                 master_netdev, file_storage_dir, candidate_pool_size,
                 secondary_ip=None, vg_name=None, beparams=None,
                 nicparams=None, hvparams=None, enabled_hypervisors=None,
-                modify_etc_hosts=True, modify_ssh_setup=True):
+                modify_etc_hosts=True, modify_ssh_setup=True,
+                maintain_node_health=False):
   """Initialise the cluster.
 
   @type candidate_pool_size: int
@@ -340,6 +341,7 @@ def InitCluster(cluster_name, mac_prefix,
     ctime=now,
     mtime=now,
     uuid=utils.NewUUID(),
+    maintain_node_health=maintain_node_health,
     )
   master_node_config = objects.Node(name=hostname.name,
                                     primary_ip=hostname.ip,

lib/cli.py

@@ -76,6 +76,7 @@ __all__ = [
   "IGNORE_SECONDARIES_OPT",
   "IGNORE_SIZE_OPT",
   "MAC_PREFIX_OPT",
+  "MAINTAIN_NODE_HEALTH_OPT",
   "MASTER_NETDEV_OPT",
   "MC_OPT",
   "NET_OPT",
@@ -913,6 +914,13 @@ USE_REPL_NET_OPT = cli_option("--use-replication-network",
                               " for talking to the nodes",
                               action="store_true", default=False)
 
+MAINTAIN_NODE_HEALTH_OPT = \
+    cli_option("--maintain-node-health", dest="maintain_node_health",
+               metavar=_YORNO, default=None, type="bool",
+               help="Configure the cluster to automatically maintain node"
+               " health, by shutting down unknown instances, shutting down"
+               " unknown DRBD devices, etc.")
+
 
 def _ParseArgs(argv, commands, aliases):
   """Parser for the command line arguments.

lib/cmdlib.py

@@ -2240,6 +2240,7 @@ class LUSetClusterParams(LogicalUnit):
       if self.op.candidate_pool_size < 1:
         raise errors.OpPrereqError("At least one master candidate needed",
                                    errors.ECODE_INVAL)
+    _CheckBooleanOpField(self.op, "maintain_node_health")
 
   def ExpandNames(self):
     # FIXME: in the future maybe other cluster params won't require checking on
@@ -2432,6 +2433,9 @@ class LUSetClusterParams(LogicalUnit):
       # we need to update the pool size here, otherwise the save will fail
       _AdjustCandidatePool(self, [])
 
+    if self.op.maintain_node_health is not None:
+      self.cluster.maintain_node_health = self.op.maintain_node_health
+
     self.cfg.Update(self.cluster, feedback_fn)
 
 
@@ -3691,6 +3695,7 @@ class LUQueryClusterInfo(NoHooksLU):
       "master_netdev": cluster.master_netdev,
       "volume_group_name": cluster.volume_group_name,
       "file_storage_dir": cluster.file_storage_dir,
+      "maintain_node_health": cluster.maintain_node_health,
       "ctime": cluster.ctime,
       "mtime": cluster.mtime,
       "uuid": cluster.uuid,

lib/objects.py

@@ -864,6 +864,7 @@ class Cluster(TaggableObject):
     "candidate_pool_size",
     "modify_etc_hosts",
     "modify_ssh_setup",
+    "maintain_node_health",
     ] + _TIMESTAMPS + _UUID
 
   def UpgradeConfig(self):
@@ -910,6 +911,10 @@ class Cluster(TaggableObject):
          if hvname != self.default_hypervisor])
       self.default_hypervisor = None
 
+    # maintain_node_health added after 2.1.1
+    if self.maintain_node_health is None:
+      self.maintain_node_health = False
+
   def ToDict(self):
     """Custom function for cluster.
 

lib/opcodes.py

@@ -305,6 +305,7 @@ class OpSetClusterParams(OpCode):
     "beparams",
     "nicparams",
     "candidate_pool_size",
+    "maintain_node_health",
     ]
 
 

man/gnt-cluster.sgml

@@ -242,6 +242,8 @@
         <sbr>
         <arg>--nic-parameters <replaceable>nic-param</replaceable>=<replaceable>value</replaceable><arg rep="repeat" choice="opt">,<replaceable>nic-param</replaceable>=<replaceable>value</replaceable></arg></arg>
         <sbr>
+        <arg>--maintain-node-health <group choice="req"><arg>yes</arg><arg>no</arg></group></arg>
+        <sbr>
         <arg choice="req"><replaceable>clustername</replaceable></arg>
       </cmdsynopsis>
 
@@ -489,6 +491,15 @@
         </variablelist>
       </para>
 
+      <para>
+        The option <option>--maintain-node-health</option> allows to
+        enable/disable automatic maintenance actions on
+        nodes. Currently these include automatic shutdown of instances
+        and deactivation of DRBD devices on offline nodes; in the
+        future it might be extended to automatic removal of unknown
+        LVM volumes, etc.
+      </para>
+
     </refsect2>
 
     <refsect2>
@@ -546,6 +557,8 @@
         <arg choice="opt">--nic-parameters <replaceable>nic-param</replaceable>=<replaceable>value</replaceable><arg rep="repeat" choice="opt">,<replaceable>nic-param</replaceable>=<replaceable>value</replaceable></arg></arg>
         <sbr>
         <arg choice="opt">-C <replaceable>candidate_pool_size</replaceable></arg>
+        <sbr>
+        <arg>--maintain-node-health <group choice="req"><arg>yes</arg><arg>no</arg></group></arg>
 
       </cmdsynopsis>
 
@@ -558,7 +571,8 @@
           <option>--enabled-hypervisors</option>,
           <option>--hypervisor-parameters</option>,
           <option>--backend-parameters</option> and
-          <option>--nic-parameters</option> and options are
+          <option>--nic-parameters</option>
+          and <option>--maintain-node-health</option> options are
           described in the <command>init</command> command.
         </para>
 

scripts/gnt-cluster

@@ -104,6 +104,7 @@ def InitCluster(opts, args):
                         candidate_pool_size=opts.candidate_pool_size,
                         modify_etc_hosts=opts.modify_etc_hosts,
                         modify_ssh_setup=opts.modify_ssh_setup,
+                        maintain_node_health=opts.maintain_node_health,
                         )
   op = opcodes.OpPostInitCluster()
   SubmitOpCode(op, opts=opts)
@@ -272,6 +273,8 @@ def ShowClusterConfig(opts, args):
   ToStdout("  - master netdev: %s", result["master_netdev"])
   ToStdout("  - lvm volume group: %s", result["volume_group_name"])
   ToStdout("  - file storage path: %s", result["file_storage_dir"])
+  ToStdout("  - maintenance of node health: %s",
+           result["maintain_node_health"])
 
   ToStdout("Default instance parameters:")
   _PrintGroupedParams(result["beparams"])
@@ -597,7 +600,8 @@ def SetClusterParams(opts, args):
   if not (not opts.lvm_storage or opts.vg_name or
           opts.enabled_hypervisors or opts.hvparams or
           opts.beparams or opts.nicparams or
-          opts.candidate_pool_size is not None):
+          opts.candidate_pool_size is not None or
+          opts.maintain_node_health is not None):
     ToStderr("Please give at least one of the parameters.")
     return 1
 
@@ -624,13 +628,16 @@ def SetClusterParams(opts, args):
   nicparams = opts.nicparams
   utils.ForceDictType(nicparams, constants.NICS_PARAMETER_TYPES)
 
+  mnh = opts.maintain_node_health
+
   op = opcodes.OpSetClusterParams(vg_name=vg_name,
                                   enabled_hypervisors=hvlist,
                                   hvparams=hvparams,
                                   os_hvp=None,
                                   beparams=beparams,
                                   nicparams=nicparams,
-                                  candidate_pool_size=opts.candidate_pool_size)
+                                  candidate_pool_size=opts.candidate_pool_size,
+                                  maintain_node_health=mnh)
   SubmitOpCode(op, opts=opts)
   return 0
 
@@ -712,7 +719,7 @@ commands = {
     [BACKEND_OPT, CP_SIZE_OPT, ENABLED_HV_OPT, GLOBAL_FILEDIR_OPT,
      HVLIST_OPT, MAC_PREFIX_OPT, MASTER_NETDEV_OPT, NIC_PARAMS_OPT,
      NOLVM_STORAGE_OPT, NOMODIFY_ETCHOSTS_OPT, NOMODIFY_SSH_SETUP_OPT,
-     SECONDARY_IP_OPT, VG_NAME_OPT],
+     SECONDARY_IP_OPT, VG_NAME_OPT, MAINTAIN_NODE_HEALTH_OPT],
     "[opts...] <cluster_name>", "Initialises a new cluster configuration"),
   'destroy': (
     DestroyCluster, ARGS_NONE, [YES_DOIT_OPT],
@@ -781,7 +788,7 @@ commands = {
   'modify': (
     SetClusterParams, ARGS_NONE,
     [BACKEND_OPT, CP_SIZE_OPT, ENABLED_HV_OPT, HVLIST_OPT,
-     NIC_PARAMS_OPT, NOLVM_STORAGE_OPT, VG_NAME_OPT],
+     NIC_PARAMS_OPT, NOLVM_STORAGE_OPT, VG_NAME_OPT, MAINTAIN_NODE_HEALTH_OPT],
     "[opts...]",
     "Alters the parameters of the cluster"),
   "renew-crypto": (