Add timeout options to other LUs

All the LUs that shut down the instance need to be able too pass the
timeout parameter as well.
Signed-off-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

daemons/ganeti-noded

@@ -502,7 +502,8 @@ class NodeHttpServer(http.server.HttpServer):
     """
     instance = objects.Instance.FromDict(params[0])
     reboot_type = params[1]
-    return backend.InstanceReboot(instance, reboot_type)
+    shutdown_timeout = params[2]
+    return backend.InstanceReboot(instance, reboot_type, shutdown_timeout)
 
   @staticmethod
   def perspective_instance_info(params):

lib/backend.py

@@ -1015,7 +1015,7 @@ def InstanceShutdown(instance, timeout):
   _RemoveBlockDevLinks(iname, instance.disks)
 
 
-def InstanceReboot(instance, reboot_type):
+def InstanceReboot(instance, reboot_type, shutdown_timeout):
   """Reboot an instance.
 
   @type instance: L{objects.Instance}
@@ -1031,6 +1031,8 @@ def InstanceReboot(instance, reboot_type):
         not accepted here, since that mode is handled differently, in
         cmdlib, and translates into full stop and start of the
         instance (instead of a call_instance_reboot RPC)
+  @type timeout: integer
+  @param timeout: maximum timeout for soft shutdown
   @rtype: None
 
   """
@@ -1047,7 +1049,7 @@ def InstanceReboot(instance, reboot_type):
       _Fail("Failed to soft reboot instance %s: %s", instance.name, err)
   elif reboot_type == constants.INSTANCE_REBOOT_HARD:
     try:
-      InstanceShutdown(instance)
+      InstanceShutdown(instance, shutdown_timeout)
       return StartInstance(instance)
     except errors.HypervisorError, err:
       _Fail("Failed to hard reboot instance %s: %s", instance.name, err)

lib/cmdlib.py

@@ -3598,6 +3598,13 @@ class LURebootInstance(LogicalUnit):
   _OP_REQP = ["instance_name", "ignore_secondaries", "reboot_type"]
   REQ_BGL = False
 
+  def CheckArguments(self):
+    """Check the arguments.
+
+    """
+    self.shutdown_timeout = getattr(self.op, "shutdown_timeout",
+                                    constants.DEFAULT_SHUTDOWN_TIMEOUT)
+
   def ExpandNames(self):
     if self.op.reboot_type not in [constants.INSTANCE_REBOOT_SOFT,
                                    constants.INSTANCE_REBOOT_HARD,
@@ -3617,6 +3624,7 @@ class LURebootInstance(LogicalUnit):
     env = {
       "IGNORE_SECONDARIES": self.op.ignore_secondaries,
       "REBOOT_TYPE": self.op.reboot_type,
+      "SHUTDOWN_TIMEOUT": self.shutdown_timeout,
       }
     env.update(_BuildInstanceHookEnvByObject(self, self.instance))
     nl = [self.cfg.GetMasterNode()] + list(self.instance.all_nodes)
@@ -3652,10 +3660,12 @@ class LURebootInstance(LogicalUnit):
       for disk in instance.disks:
         self.cfg.SetDiskID(disk, node_current)
       result = self.rpc.call_instance_reboot(node_current, instance,
-                                             reboot_type)
+                                             reboot_type,
+                                             self.shutdown_timeout)
       result.Raise("Could not reboot instance")
     else:
-      result = self.rpc.call_instance_shutdown(node_current, instance)
+      result = self.rpc.call_instance_shutdown(node_current, instance,
+                                               self.shutdown_timeout)
       result.Raise("Could not shutdown instance for full reboot")
       _ShutdownInstanceDisks(self, instance)
       _StartInstanceDisks(self, instance, ignore_secondaries)
@@ -4008,6 +4018,13 @@ class LURemoveInstance(LogicalUnit):
   _OP_REQP = ["instance_name", "ignore_failures"]
   REQ_BGL = False
 
+  def CheckArguments(self):
+    """Check the arguments.
+
+    """
+    self.shutdown_timeout = getattr(self.op, "shutdown_timeout",
+                                    constants.DEFAULT_SHUTDOWN_TIMEOUT)
+
   def ExpandNames(self):
     self._ExpandAndLockInstance()
     self.needed_locks[locking.LEVEL_NODE] = []
@@ -4024,6 +4041,7 @@ class LURemoveInstance(LogicalUnit):
 
     """
     env = _BuildInstanceHookEnvByObject(self, self.instance)
+    env["SHUTDOWN_TIMEOUT"] = self.shutdown_timeout
     nl = [self.cfg.GetMasterNode()]
     return env, nl, nl
 
@@ -4045,7 +4063,8 @@ class LURemoveInstance(LogicalUnit):
     logging.info("Shutting down instance %s on node %s",
                  instance.name, instance.primary_node)
 
-    result = self.rpc.call_instance_shutdown(instance.primary_node, instance)
+    result = self.rpc.call_instance_shutdown(instance.primary_node, instance,
+                                             self.shutdown_timeout)
     msg = result.fail_msg
     if msg:
       if self.op.ignore_failures:
@@ -4356,6 +4375,13 @@ class LUFailoverInstance(LogicalUnit):
   _OP_REQP = ["instance_name", "ignore_consistency"]
   REQ_BGL = False
 
+  def CheckArguments(self):
+    """Check the arguments.
+
+    """
+    self.shutdown_timeout = getattr(self.op, "shutdown_timeout",
+                                    constants.DEFAULT_SHUTDOWN_TIMEOUT)
+
   def ExpandNames(self):
     self._ExpandAndLockInstance()
     self.needed_locks[locking.LEVEL_NODE] = []
@@ -4373,6 +4399,7 @@ class LUFailoverInstance(LogicalUnit):
     """
     env = {
       "IGNORE_CONSISTENCY": self.op.ignore_consistency,
+      "SHUTDOWN_TIMEOUT": self.shutdown_timeout,
       }
     env.update(_BuildInstanceHookEnvByObject(self, self.instance))
     nl = [self.cfg.GetMasterNode()] + list(self.instance.secondary_nodes)
@@ -4437,7 +4464,8 @@ class LUFailoverInstance(LogicalUnit):
     logging.info("Shutting down instance %s on node %s",
                  instance.name, source_node)
 
-    result = self.rpc.call_instance_shutdown(source_node, instance)
+    result = self.rpc.call_instance_shutdown(source_node, instance,
+                                             self.shutdown_timeout)
     msg = result.fail_msg
     if msg:
       if self.op.ignore_consistency:
@@ -4529,6 +4557,13 @@ class LUMoveInstance(LogicalUnit):
   _OP_REQP = ["instance_name", "target_node"]
   REQ_BGL = False
 
+  def CheckArguments(self):
+    """Check the arguments.
+
+    """
+    self.shutdown_timeout = getattr(self.op, "shutdown_timeout",
+                                    constants.DEFAULT_SHUTDOWN_TIMEOUT)
+
   def ExpandNames(self):
     self._ExpandAndLockInstance()
     target_node = self.cfg.ExpandNodeName(self.op.target_node)
@@ -4551,6 +4586,7 @@ class LUMoveInstance(LogicalUnit):
     """
     env = {
       "TARGET_NODE": self.op.target_node,
+      "SHUTDOWN_TIMEOUT": self.shutdown_timeout,
       }
     env.update(_BuildInstanceHookEnvByObject(self, self.instance))
     nl = [self.cfg.GetMasterNode()] + [self.instance.primary_node,
@@ -4614,7 +4650,8 @@ class LUMoveInstance(LogicalUnit):
     self.LogInfo("Shutting down instance %s on source node %s",
                  instance.name, source_node)
 
-    result = self.rpc.call_instance_shutdown(source_node, instance)
+    result = self.rpc.call_instance_shutdown(source_node, instance,
+                                             self.shutdown_timeout)
     msg = result.fail_msg
     if msg:
       if self.op.ignore_consistency:
@@ -7687,6 +7724,13 @@ class LUExportInstance(LogicalUnit):
   _OP_REQP = ["instance_name", "target_node", "shutdown"]
   REQ_BGL = False
 
+  def CheckArguments(self):
+    """Check the arguments.
+
+    """
+    self.shutdown_timeout = getattr(self.op, "shutdown_timeout",
+                                    constants.DEFAULT_SHUTDOWN_TIMEOUT)
+
   def ExpandNames(self):
     self._ExpandAndLockInstance()
     # FIXME: lock only instance primary and destination node
@@ -7712,6 +7756,7 @@ class LUExportInstance(LogicalUnit):
     env = {
       "EXPORT_NODE": self.op.target_node,
       "EXPORT_DO_SHUTDOWN": self.op.shutdown,
+      "SHUTDOWN_TIMEOUT": self.shutdown_timeout,
       }
     env.update(_BuildInstanceHookEnvByObject(self, self.instance))
     nl = [self.cfg.GetMasterNode(), self.instance.primary_node,
@@ -7756,7 +7801,8 @@ class LUExportInstance(LogicalUnit):
     if self.op.shutdown:
       # shutdown the instance, but not the disks
       feedback_fn("Shutting down instance %s" % instance.name)
-      result = self.rpc.call_instance_shutdown(src_node, instance)
+      result = self.rpc.call_instance_shutdown(src_node, instance,
+                                               self.shutdown_timeout)
       result.Raise("Could not shutdown instance %s on"
                    " node %s" % (instance.name, src_node))
 

lib/opcodes.py

@@ -497,7 +497,7 @@ class OpRebootInstance(OpCode):
   OP_ID = "OP_INSTANCE_REBOOT"
   OP_DSC_FIELD = "instance_name"
   __slots__ = OpCode.__slots__ + [
-    "instance_name", "reboot_type", "ignore_secondaries",
+    "instance_name", "reboot_type", "ignore_secondaries", "shutdown_timeout",
     ]
 
 
@@ -514,7 +514,9 @@ class OpFailoverInstance(OpCode):
   """Failover an instance."""
   OP_ID = "OP_INSTANCE_FAILOVER"
   OP_DSC_FIELD = "instance_name"
-  __slots__ = OpCode.__slots__ + ["instance_name", "ignore_consistency"]
+  __slots__ = OpCode.__slots__ + [
+    "instance_name", "ignore_consistency", "shutdown_timeout",
+    ]
 
 
 class OpMigrateInstance(OpCode):
@@ -543,7 +545,9 @@ class OpMoveInstance(OpCode):
   """
   OP_ID = "OP_INSTANCE_MOVE"
   OP_DSC_FIELD = "instance_name"
-  __slots__ = OpCode.__slots__ + ["instance_name", "target_node"]
+  __slots__ = OpCode.__slots__ + [
+    "instance_name", "target_node", "shutdown_timeout",
+  ]
 
 
 class OpConnectConsole(OpCode):
@@ -624,7 +628,9 @@ class OpExportInstance(OpCode):
   """Export an instance."""
   OP_ID = "OP_BACKUP_EXPORT"
   OP_DSC_FIELD = "instance_name"
-  __slots__ = OpCode.__slots__ + ["instance_name", "target_node", "shutdown"]
+  __slots__ = OpCode.__slots__ + [
+    "instance_name", "target_node", "shutdown", "shutdown_timeout",
+    ]
 
 
 class OpRemoveExport(OpCode):

lib/rpc.py

@@ -548,14 +548,15 @@ class RpcRunner(object):
     return self._SingleNodeCall(node, "instance_migrate",
                                 [self._InstDict(instance), target, live])
 
-  def call_instance_reboot(self, node, instance, reboot_type):
+  def call_instance_reboot(self, node, inst, reboot_type, shutdown_timeout):
     """Reboots an instance.
 
     This is a single-node call.
 
     """
     return self._SingleNodeCall(node, "instance_reboot",
-                                [self._InstDict(instance), reboot_type])
+                                [self._InstDict(inst), reboot_type,
+                                 shutdown_timeout])
 
   def call_instance_os_add(self, node, inst, reinstall):
     """Installs an OS on the given instance.

scripts/gnt-backup

@@ -72,7 +72,8 @@ def ExportInstance(opts, args):
   """
   op = opcodes.OpExportInstance(instance_name=args[0],
                                 target_node=opts.node,
-                                shutdown=opts.shutdown)
+                                shutdown=opts.shutdown,
+                                timeout=opts.timeout)
 
   fin_resu, dlist = SubmitOpCode(op)
   if not isinstance(dlist, list):
@@ -150,7 +151,7 @@ commands = {
     "", "Lists instance exports available in the ganeti cluster"),
   'export': (
     ExportInstance, ARGS_ONE_INSTANCE,
-    [FORCE_OPT, SINGLE_NODE_OPT, NOSHUTDOWN_OPT],
+    [FORCE_OPT, SINGLE_NODE_OPT, NOSHUTDOWN_OPT, SHUTDOWN_TIMEOUT_OPT],
     "-n <target_node> [opts...] <name>",
     "Exports an instance to an image"),
   'import': (

scripts/gnt-instance

@@ -563,7 +563,8 @@ def RemoveInstance(opts, args):
       return 1
 
   op = opcodes.OpRemoveInstance(instance_name=instance_name,
-                                ignore_failures=opts.ignore_failures)
+                                ignore_failures=opts.ignore_failures,
+                                timeout=opts.timeout)
   SubmitOrSend(op, opts, cl=cl)
   return 0
 
@@ -713,7 +714,8 @@ def _RebootInstance(name, opts):
   """
   return opcodes.OpRebootInstance(instance_name=name,
                                   reboot_type=opts.reboot_type,
-                                  ignore_secondaries=opts.ignore_secondaries)
+                                  ignore_secondaries=opts.ignore_secondaries,
+                                  timeout=opts.timeout)
 
 
 def _ShutdownInstance(name, opts):
@@ -803,7 +805,8 @@ def FailoverInstance(opts, args):
       return 1
 
   op = opcodes.OpFailoverInstance(instance_name=instance_name,
-                                  ignore_consistency=opts.ignore_consistency)
+                                  ignore_consistency=opts.ignore_consistency,
+                                  timeout=opts.timeout)
   SubmitOrSend(op, opts, cl=cl)
   return 0
 
@@ -868,7 +871,8 @@ def MoveInstance(opts, args):
       return 1
 
   op = opcodes.OpMoveInstance(instance_name=instance_name,
-                              target_node=opts.node)
+                              target_node=opts.node,
+                              timeout=opts.timeout)
   SubmitOrSend(op, opts, cl=cl)
   return 0
 
@@ -1289,7 +1293,7 @@ commands = {
     "[--show-cmd] <instance>", "Opens a console on the specified instance"),
   'failover': (
     FailoverInstance, ARGS_ONE_INSTANCE,
-    [FORCE_OPT, IGNORE_CONSIST_OPT, SUBMIT_OPT],
+    [FORCE_OPT, IGNORE_CONSIST_OPT, SUBMIT_OPT, SHUTDOWN_TIMEOUT_OPT],
     "[-f] <instance>", "Stops the instance and starts it on the backup node,"
     " using the remote mirror (only for instances of type drbd)"),
   'migrate': (
@@ -1299,7 +1303,7 @@ commands = {
     " (only for instances of type drbd)"),
   'move': (
     MoveInstance, ARGS_ONE_INSTANCE,
-    [FORCE_OPT, SUBMIT_OPT, SINGLE_NODE_OPT],
+    [FORCE_OPT, SUBMIT_OPT, SINGLE_NODE_OPT, SHUTDOWN_TIMEOUT_OPT],
     "[-f] <instance>", "Move instance to an arbitrary node"
     " (only for instances of type file and lv)"),
   'info': (
@@ -1327,7 +1331,7 @@ commands = {
     "[-f] <instance>", "Reinstall a stopped instance"),
   'remove': (
     RemoveInstance, ARGS_ONE_INSTANCE,
-    [FORCE_OPT, IGNORE_FAILURES_OPT, SUBMIT_OPT],
+    [FORCE_OPT, SHUTDOWN_TIMEOUT_OPT, IGNORE_FAILURES_OPT, SUBMIT_OPT],
     "[-f] <instance>", "Shuts down the instance and removes it"),
   'rename': (
     RenameInstance,
@@ -1358,7 +1362,8 @@ commands = {
   'reboot': (
     GenericManyOps("reboot", _RebootInstance), [ArgInstance()],
     [m_force_multi, REBOOT_TYPE_OPT, IGNORE_SECONDARIES_OPT, m_node_opt,
-     m_pri_node_opt, m_sec_node_opt, m_clust_opt, m_inst_opt, SUBMIT_OPT],
+     m_pri_node_opt, m_sec_node_opt, m_clust_opt, m_inst_opt, SUBMIT_OPT,
+     SHUTDOWN_TIMEOUT_OPT],
     "<instance>", "Reboots an instance"),
   'activate-disks': (
     ActivateDisks, ARGS_ONE_INSTANCE, [SUBMIT_OPT, IGNORE_SIZE_OPT],