Improve LUQueryNodes for lockless case
[ganeti-local] / lib / cmdlib.py
index 0e2e184..87ce76d 100644 (file)
 
 import os
 import os.path
 
 import os
 import os.path
-import sha
 import time
 import time
-import tempfile
 import re
 import platform
 import logging
 import copy
 import re
 import platform
 import logging
 import copy
-import random
 
 from ganeti import ssh
 from ganeti import utils
 
 from ganeti import ssh
 from ganeti import utils
@@ -41,7 +38,6 @@ from ganeti import hypervisor
 from ganeti import locking
 from ganeti import constants
 from ganeti import objects
 from ganeti import locking
 from ganeti import constants
 from ganeti import objects
-from ganeti import opcodes
 from ganeti import serializer
 from ganeti import ssconf
 
 from ganeti import serializer
 from ganeti import ssconf
 
@@ -69,7 +65,7 @@ class LogicalUnit(object):
   def __init__(self, processor, op, context, rpc):
     """Constructor for LogicalUnit.
 
   def __init__(self, processor, op, context, rpc):
     """Constructor for LogicalUnit.
 
-    This needs to be overriden in derived classes in order to check op
+    This needs to be overridden in derived classes in order to check op
     validity.
 
     """
     validity.
 
     """
@@ -117,7 +113,7 @@ class LogicalUnit(object):
     CheckPrereq, doing these separate is better because:
 
       - ExpandNames is left as as purely a lock-related function
     CheckPrereq, doing these separate is better because:
 
       - ExpandNames is left as as purely a lock-related function
-      - CheckPrereq is run after we have aquired locks (and possible
+      - CheckPrereq is run after we have acquired locks (and possible
         waited for them)
 
     The function is allowed to change the self.op attribute so that
         waited for them)
 
     The function is allowed to change the self.op attribute so that
@@ -454,7 +450,8 @@ def _CheckNodeNotDrained(lu, node):
 
 
 def _BuildInstanceHookEnv(name, primary_node, secondary_nodes, os_type, status,
 
 
 def _BuildInstanceHookEnv(name, primary_node, secondary_nodes, os_type, status,
-                          memory, vcpus, nics, disk_template, disks):
+                          memory, vcpus, nics, disk_template, disks,
+                          bep, hvp, hypervisor_name):
   """Builds instance related env variables for hooks
 
   This builds the hook environment from individual variables.
   """Builds instance related env variables for hooks
 
   This builds the hook environment from individual variables.
@@ -477,9 +474,15 @@ def _BuildInstanceHookEnv(name, primary_node, secondary_nodes, os_type, status,
   @param nics: list of tuples (ip, bridge, mac) representing
       the NICs the instance  has
   @type disk_template: string
   @param nics: list of tuples (ip, bridge, mac) representing
       the NICs the instance  has
   @type disk_template: string
-  @param disk_template: the distk template of the instance
+  @param disk_template: the disk template of the instance
   @type disks: list
   @param disks: the list of (size, mode) pairs
   @type disks: list
   @param disks: the list of (size, mode) pairs
+  @type bep: dict
+  @param bep: the backend parameters for the instance
+  @type hvp: dict
+  @param hvp: the hypervisor parameters for the instance
+  @type hypervisor_name: string
+  @param hypervisor_name: the hypervisor for the instance
   @rtype: dict
   @return: the hook environment for this instance
 
   @rtype: dict
   @return: the hook environment for this instance
 
@@ -498,6 +501,7 @@ def _BuildInstanceHookEnv(name, primary_node, secondary_nodes, os_type, status,
     "INSTANCE_MEMORY": memory,
     "INSTANCE_VCPUS": vcpus,
     "INSTANCE_DISK_TEMPLATE": disk_template,
     "INSTANCE_MEMORY": memory,
     "INSTANCE_VCPUS": vcpus,
     "INSTANCE_DISK_TEMPLATE": disk_template,
+    "INSTANCE_HYPERVISOR": hypervisor_name,
   }
 
   if nics:
   }
 
   if nics:
@@ -523,6 +527,10 @@ def _BuildInstanceHookEnv(name, primary_node, secondary_nodes, os_type, status,
 
   env["INSTANCE_DISK_COUNT"] = disk_count
 
 
   env["INSTANCE_DISK_COUNT"] = disk_count
 
+  for source, kind in [(bep, "BE"), (hvp, "HV")]:
+    for key, value in source.items():
+      env["INSTANCE_%s_%s" % (kind, key)] = value
+
   return env
 
 
   return env
 
 
@@ -541,7 +549,9 @@ def _BuildInstanceHookEnvByObject(lu, instance, override=None):
   @return: the hook environment dictionary
 
   """
   @return: the hook environment dictionary
 
   """
-  bep = lu.cfg.GetClusterInfo().FillBE(instance)
+  cluster = lu.cfg.GetClusterInfo()
+  bep = cluster.FillBE(instance)
+  hvp = cluster.FillHV(instance)
   args = {
     'name': instance.name,
     'primary_node': instance.primary_node,
   args = {
     'name': instance.name,
     'primary_node': instance.primary_node,
@@ -553,6 +563,9 @@ def _BuildInstanceHookEnvByObject(lu, instance, override=None):
     'nics': [(nic.ip, nic.bridge, nic.mac) for nic in instance.nics],
     'disk_template': instance.disk_template,
     'disks': [(disk.size, disk.mode) for disk in instance.disks],
     'nics': [(nic.ip, nic.bridge, nic.mac) for nic in instance.nics],
     'disk_template': instance.disk_template,
     'disks': [(disk.size, disk.mode) for disk in instance.disks],
+    'bep': bep,
+    'hvp': hvp,
+    'hypervisor_name': instance.hypervisor,
   }
   if override:
     args.update(override)
   }
   if override:
     args.update(override)
@@ -576,10 +589,10 @@ def _AdjustCandidatePool(lu):
 
 
 def _CheckInstanceBridgesExist(lu, instance):
 
 
 def _CheckInstanceBridgesExist(lu, instance):
-  """Check that the brigdes needed by an instance exist.
+  """Check that the bridges needed by an instance exist.
 
   """
 
   """
-  # check bridges existance
+  # check bridges existence
   brlist = [nic.bridge for nic in instance.nics]
   result = lu.rpc.call_bridges_exist(instance.primary_node, brlist)
   result.Raise()
   brlist = [nic.bridge for nic in instance.nics]
   result = lu.rpc.call_bridges_exist(instance.primary_node, brlist)
   result.Raise()
@@ -600,7 +613,7 @@ class LUDestroyCluster(NoHooksLU):
 
     This checks whether the cluster is empty.
 
 
     This checks whether the cluster is empty.
 
-    Any errors are signalled by raising errors.OpPrereqError.
+    Any errors are signaled by raising errors.OpPrereqError.
 
     """
     master = self.cfg.GetMasterNode()
 
     """
     master = self.cfg.GetMasterNode()
@@ -653,7 +666,7 @@ class LUVerifyCluster(LogicalUnit):
     Test list:
 
       - compares ganeti version
     Test list:
 
       - compares ganeti version
-      - checks vg existance and size > 20G
+      - checks vg existence and size > 20G
       - checks config file checksum
       - checks ssh to other nodes
 
       - checks config file checksum
       - checks ssh to other nodes
 
@@ -735,8 +748,8 @@ class LUVerifyCluster(LogicalUnit):
           else:
             # not candidate and this is not a must-have file
             bad = True
           else:
             # not candidate and this is not a must-have file
             bad = True
-            feedback_fn("  - ERROR: non master-candidate has old/wrong file"
-                        " '%s'" % file_name)
+            feedback_fn("  - ERROR: file '%s' should not exist on non master"
+                        " candidates (and the file is outdated)" % file_name)
         else:
           # all good, except non-master/non-must have combination
           if not node_is_mc and not must_have_file:
         else:
           # all good, except non-master/non-must have combination
           if not node_is_mc and not must_have_file:
@@ -892,7 +905,7 @@ class LUVerifyCluster(LogicalUnit):
           if bep[constants.BE_AUTO_BALANCE]:
             needed_mem += bep[constants.BE_MEMORY]
         if nodeinfo['mfree'] < needed_mem:
           if bep[constants.BE_AUTO_BALANCE]:
             needed_mem += bep[constants.BE_MEMORY]
         if nodeinfo['mfree'] < needed_mem:
-          feedback_fn("  - ERROR: not enough memory on node %s to accomodate"
+          feedback_fn("  - ERROR: not enough memory on node %s to accommodate"
                       " failovers should node %s fail" % (node, prinode))
           bad = True
     return bad
                       " failovers should node %s fail" % (node, prinode))
           bad = True
     return bad
@@ -911,7 +924,7 @@ class LUVerifyCluster(LogicalUnit):
   def BuildHooksEnv(self):
     """Build hooks env.
 
   def BuildHooksEnv(self):
     """Build hooks env.
 
-    Cluster-Verify hooks just rone in the post phase and their failure makes
+    Cluster-Verify hooks just ran in the post phase and their failure makes
     the output be logged in the verify output and the verification to fail.
 
     """
     the output be logged in the verify output and the verification to fail.
 
     """
@@ -1178,7 +1191,7 @@ class LUVerifyCluster(LogicalUnit):
     return not bad
 
   def HooksCallBack(self, phase, hooks_results, feedback_fn, lu_result):
     return not bad
 
   def HooksCallBack(self, phase, hooks_results, feedback_fn, lu_result):
-    """Analize the post-hooks' result
+    """Analyze the post-hooks' result
 
     This method analyses the hook result, handles it, and sends some
     nicely-formatted feedback back to the user.
 
     This method analyses the hook result, handles it, and sends some
     nicely-formatted feedback back to the user.
@@ -1277,7 +1290,6 @@ class LUVerifyDisks(NoHooksLU):
 
     node_lvs = self.rpc.call_volume_list(nodes, vg_name)
 
 
     node_lvs = self.rpc.call_volume_list(nodes, vg_name)
 
-    to_act = set()
     for node in nodes:
       # node_volume
       lvs = node_lvs[node]
     for node in nodes:
       # node_volume
       lvs = node_lvs[node]
@@ -1313,6 +1325,128 @@ class LUVerifyDisks(NoHooksLU):
     return result
 
 
     return result
 
 
+class LURepairDiskSizes(NoHooksLU):
+  """Verifies the cluster disks sizes.
+
+  """
+  _OP_REQP = ["instances"]
+  REQ_BGL = False
+
+  def ExpandNames(self):
+
+    if not isinstance(self.op.instances, list):
+      raise errors.OpPrereqError("Invalid argument type 'instances'")
+
+    if self.op.instances:
+      self.wanted_names = []
+      for name in self.op.instances:
+        full_name = self.cfg.ExpandInstanceName(name)
+        if full_name is None:
+          raise errors.OpPrereqError("Instance '%s' not known" % name)
+        self.wanted_names.append(full_name)
+      self.needed_locks = {
+        locking.LEVEL_NODE: [],
+        locking.LEVEL_INSTANCE: self.wanted_names,
+        }
+      self.recalculate_locks[locking.LEVEL_NODE] = constants.LOCKS_REPLACE
+    else:
+      self.wanted_names = None
+      self.needed_locks = {
+        locking.LEVEL_NODE: locking.ALL_SET,
+        locking.LEVEL_INSTANCE: locking.ALL_SET,
+        }
+    self.share_locks = dict(((i, 1) for i in locking.LEVELS))
+
+  def DeclareLocks(self, level):
+    if level == locking.LEVEL_NODE and self.wanted_names is not None:
+      self._LockInstancesNodes(primary_only=True)
+
+  def CheckPrereq(self):
+    """Check prerequisites.
+
+    This only checks the optional instance list against the existing names.
+
+    """
+    if self.wanted_names is None:
+      self.wanted_names = self.acquired_locks[locking.LEVEL_INSTANCE]
+
+    self.wanted_instances = [self.cfg.GetInstanceInfo(name) for name
+                             in self.wanted_names]
+
+  def _EnsureChildSizes(self, disk):
+    """Ensure children of the disk have the needed disk size.
+
+    This is valid mainly for DRBD8 and fixes an issue where the
+    children have smaller disk size.
+
+    @param disk: an L{ganeti.objects.Disk} object
+
+    """
+    if disk.dev_type == constants.LD_DRBD8:
+      assert disk.children, "Empty children for DRBD8?"
+      fchild = disk.children[0]
+      mismatch = fchild.size < disk.size
+      if mismatch:
+        self.LogInfo("Child disk has size %d, parent %d, fixing",
+                     fchild.size, disk.size)
+        fchild.size = disk.size
+
+      # and we recurse on this child only, not on the metadev
+      return self._EnsureChildSizes(fchild) or mismatch
+    else:
+      return False
+
+  def Exec(self, feedback_fn):
+    """Verify the size of cluster disks.
+
+    """
+    # TODO: check child disks too
+    # TODO: check differences in size between primary/secondary nodes
+    per_node_disks = {}
+    for instance in self.wanted_instances:
+      pnode = instance.primary_node
+      if pnode not in per_node_disks:
+        per_node_disks[pnode] = []
+      for idx, disk in enumerate(instance.disks):
+        per_node_disks[pnode].append((instance, idx, disk))
+
+    changed = []
+    for node, dskl in per_node_disks.items():
+      newl = [v[2].Copy() for v in dskl]
+      for dsk in newl:
+        self.cfg.SetDiskID(dsk, node)
+      result = self.rpc.call_blockdev_getsizes(node, newl)
+      if result.failed:
+        self.LogWarning("Failure in blockdev_getsizes call to node"
+                        " %s, ignoring", node)
+        continue
+      if len(result.data) != len(dskl):
+        self.LogWarning("Invalid result from node %s, ignoring node results",
+                        node)
+        continue
+      for ((instance, idx, disk), size) in zip(dskl, result.data):
+        if size is None:
+          self.LogWarning("Disk %d of instance %s did not return size"
+                          " information, ignoring", idx, instance.name)
+          continue
+        if not isinstance(size, (int, long)):
+          self.LogWarning("Disk %d of instance %s did not return valid"
+                          " size information, ignoring", idx, instance.name)
+          continue
+        size = size >> 20
+        if size != disk.size:
+          self.LogInfo("Disk %d of instance %s has mismatched size,"
+                       " correcting: recorded %d, actual %d", idx,
+                       instance.name, disk.size, size)
+          disk.size = size
+          self.cfg.Update(instance)
+          changed.append((instance.name, idx, size))
+        if self._EnsureChildSizes(disk):
+          self.cfg.Update(instance)
+          changed.append((instance.name, idx, disk.size))
+    return changed
+
+
 class LURenameCluster(LogicalUnit):
   """Rename the cluster.
 
 class LURenameCluster(LogicalUnit):
   """Rename the cluster.
 
@@ -1387,7 +1521,7 @@ class LURenameCluster(LogicalUnit):
                         constants.SSH_KNOWN_HOSTS_FILE, to_node)
 
     finally:
                         constants.SSH_KNOWN_HOSTS_FILE, to_node)
 
     finally:
-      result = self.rpc.call_node_start_master(master, False)
+      result = self.rpc.call_node_start_master(master, False, False)
       if result.failed or not result.data:
         self.LogWarning("Could not re-enable the master role on"
                         " the master, please restart manually.")
       if result.failed or not result.data:
         self.LogWarning("Could not re-enable the master role on"
                         " the master, please restart manually.")
@@ -1398,7 +1532,7 @@ def _RecursiveCheckIfLVMBased(disk):
 
   @type disk: L{objects.Disk}
   @param disk: the disk to check
 
   @type disk: L{objects.Disk}
   @param disk: the disk to check
-  @rtype: booleean
+  @rtype: boolean
   @return: boolean indicating whether a LD_LV dev_type was found or not
 
   """
   @return: boolean indicating whether a LD_LV dev_type was found or not
 
   """
@@ -1418,7 +1552,7 @@ class LUSetClusterParams(LogicalUnit):
   _OP_REQP = []
   REQ_BGL = False
 
   _OP_REQP = []
   REQ_BGL = False
 
-  def CheckParameters(self):
+  def CheckArguments(self):
     """Check parameters
 
     """
     """Check parameters
 
     """
@@ -1427,7 +1561,7 @@ class LUSetClusterParams(LogicalUnit):
     if self.op.candidate_pool_size is not None:
       try:
         self.op.candidate_pool_size = int(self.op.candidate_pool_size)
     if self.op.candidate_pool_size is not None:
       try:
         self.op.candidate_pool_size = int(self.op.candidate_pool_size)
-      except ValueError, err:
+      except (ValueError, TypeError), err:
         raise errors.OpPrereqError("Invalid candidate_pool_size value: %s" %
                                    str(err))
       if self.op.candidate_pool_size < 1:
         raise errors.OpPrereqError("Invalid candidate_pool_size value: %s" %
                                    str(err))
       if self.op.candidate_pool_size < 1:
@@ -1504,6 +1638,13 @@ class LUSetClusterParams(LogicalUnit):
 
     if self.op.enabled_hypervisors is not None:
       self.hv_list = self.op.enabled_hypervisors
 
     if self.op.enabled_hypervisors is not None:
       self.hv_list = self.op.enabled_hypervisors
+      if not self.hv_list:
+        raise errors.OpPrereqError("Enabled hypervisors list must contain at"
+                                   " least one member")
+      invalid_hvs = set(self.hv_list) - constants.HYPER_TYPES
+      if invalid_hvs:
+        raise errors.OpPrereqError("Enabled hypervisors contains invalid"
+                                   " entries: %s" % invalid_hvs)
     else:
       self.hv_list = cluster.enabled_hypervisors
 
     else:
       self.hv_list = cluster.enabled_hypervisors
 
@@ -1524,8 +1665,11 @@ class LUSetClusterParams(LogicalUnit):
 
     """
     if self.op.vg_name is not None:
 
     """
     if self.op.vg_name is not None:
-      if self.op.vg_name != self.cfg.GetVGName():
-        self.cfg.SetVGName(self.op.vg_name)
+      new_volume = self.op.vg_name
+      if not new_volume:
+        new_volume = None
+      if new_volume != self.cfg.GetVGName():
+        self.cfg.SetVGName(new_volume)
       else:
         feedback_fn("Cluster LVM configuration already in desired"
                     " state, not changing")
       else:
         feedback_fn("Cluster LVM configuration already in desired"
                     " state, not changing")
@@ -1537,14 +1681,11 @@ class LUSetClusterParams(LogicalUnit):
       self.cluster.beparams[constants.BEGR_DEFAULT] = self.new_beparams
     if self.op.candidate_pool_size is not None:
       self.cluster.candidate_pool_size = self.op.candidate_pool_size
       self.cluster.beparams[constants.BEGR_DEFAULT] = self.new_beparams
     if self.op.candidate_pool_size is not None:
       self.cluster.candidate_pool_size = self.op.candidate_pool_size
+      # we need to update the pool size here, otherwise the save will fail
+      _AdjustCandidatePool(self)
 
     self.cfg.Update(self.cluster)
 
 
     self.cfg.Update(self.cluster)
 
-    # we want to update nodes after the cluster so that if any errors
-    # happen, we have recorded and saved the cluster info
-    if self.op.candidate_pool_size is not None:
-      _AdjustCandidatePool(self)
-
 
 class LURedistributeConfig(NoHooksLU):
   """Force the redistribution of cluster configuration.
 
 class LURedistributeConfig(NoHooksLU):
   """Force the redistribution of cluster configuration.
@@ -1589,6 +1730,7 @@ def _WaitForSync(lu, instance, oneshot=False, unlock=False):
     lu.cfg.SetDiskID(dev, node)
 
   retries = 0
     lu.cfg.SetDiskID(dev, node)
 
   retries = 0
+  degr_retries = 10 # in seconds, as we sleep 1 second each time
   while True:
     max_time = 0
     done = True
   while True:
     max_time = 0
     done = True
@@ -1621,6 +1763,16 @@ def _WaitForSync(lu, instance, oneshot=False, unlock=False):
           rem_time = "no time estimate"
         lu.proc.LogInfo("- device %s: %5.2f%% done, %s" %
                         (instance.disks[i].iv_name, perc_done, rem_time))
           rem_time = "no time estimate"
         lu.proc.LogInfo("- device %s: %5.2f%% done, %s" %
                         (instance.disks[i].iv_name, perc_done, rem_time))
+
+    # if we're done but degraded, let's do a few small retries, to
+    # make sure we see a stable and not transient situation; therefore
+    # we force restart of the loop
+    if (done or oneshot) and cumul_degraded and degr_retries > 0:
+      logging.info("Degraded disks found, %d retries left", degr_retries)
+      degr_retries -= 1
+      time.sleep(1)
+      continue
+
     if done or oneshot:
       break
 
     if done or oneshot:
       break
 
@@ -1788,7 +1940,7 @@ class LURemoveNode(LogicalUnit):
      - it does not have primary or secondary instances
      - it's not the master
 
      - it does not have primary or secondary instances
      - it's not the master
 
-    Any errors are signalled by raising errors.OpPrereqError.
+    Any errors are signaled by raising errors.OpPrereqError.
 
     """
     node = self.cfg.GetNodeInfo(self.cfg.ExpandNodeName(self.op.node_name))
 
     """
     node = self.cfg.GetNodeInfo(self.cfg.ExpandNodeName(self.op.node_name))
@@ -1848,6 +2000,7 @@ class LUQueryNodes(NoHooksLU):
     "master",
     "offline",
     "drained",
     "master",
     "offline",
     "drained",
+    "role",
     )
 
   def ExpandNames(self):
     )
 
   def ExpandNames(self):
@@ -1930,10 +2083,9 @@ class LUQueryNodes(NoHooksLU):
     inst_fields = frozenset(("pinst_cnt", "pinst_list",
                              "sinst_cnt", "sinst_list"))
     if inst_fields & frozenset(self.op.output_fields):
     inst_fields = frozenset(("pinst_cnt", "pinst_list",
                              "sinst_cnt", "sinst_list"))
     if inst_fields & frozenset(self.op.output_fields):
-      instancelist = self.cfg.GetInstanceList()
+      inst_data = self.cfg.GetAllInstancesInfo()
 
 
-      for instance_name in instancelist:
-        inst = self.cfg.GetInstanceInfo(instance_name)
+      for instance_name, inst in inst_data.items():
         if inst.primary_node in node_to_primary:
           node_to_primary[inst.primary_node].add(inst.name)
         for secnode in inst.secondary_nodes:
         if inst.primary_node in node_to_primary:
           node_to_primary[inst.primary_node].add(inst.name)
         for secnode in inst.secondary_nodes:
@@ -1976,6 +2128,17 @@ class LUQueryNodes(NoHooksLU):
           val = node.drained
         elif self._FIELDS_DYNAMIC.Matches(field):
           val = live_data[node.name].get(field, None)
           val = node.drained
         elif self._FIELDS_DYNAMIC.Matches(field):
           val = live_data[node.name].get(field, None)
+        elif field == "role":
+          if node.name == master_node:
+            val = "M"
+          elif node.master_candidate:
+            val = "C"
+          elif node.drained:
+            val = "D"
+          elif node.offline:
+            val = "O"
+          else:
+            val = "R"
         else:
           raise errors.ParameterError(field)
         node_output.append(val)
         else:
           raise errors.ParameterError(field)
         node_output.append(val)
@@ -2097,7 +2260,7 @@ class LUAddNode(LogicalUnit):
      - it is resolvable
      - its parameters (single/dual homed) matches the cluster
 
      - it is resolvable
      - its parameters (single/dual homed) matches the cluster
 
-    Any errors are signalled by raising errors.OpPrereqError.
+    Any errors are signaled by raising errors.OpPrereqError.
 
     """
     node_name = self.op.node_name
 
     """
     node_name = self.op.node_name
@@ -2151,7 +2314,7 @@ class LUAddNode(LogicalUnit):
         raise errors.OpPrereqError("The master has a private ip but the"
                                    " new node doesn't have one")
 
         raise errors.OpPrereqError("The master has a private ip but the"
                                    " new node doesn't have one")
 
-    # checks reachablity
+    # checks reachability
     if not utils.TcpPing(primary_ip, constants.DEFAULT_NODED_PORT):
       raise errors.OpPrereqError("Node not reachable by ping")
 
     if not utils.TcpPing(primary_ip, constants.DEFAULT_NODED_PORT):
       raise errors.OpPrereqError("Node not reachable by ping")
 
@@ -2163,14 +2326,24 @@ class LUAddNode(LogicalUnit):
                                    " based ping to noded port")
 
     cp_size = self.cfg.GetClusterInfo().candidate_pool_size
                                    " based ping to noded port")
 
     cp_size = self.cfg.GetClusterInfo().candidate_pool_size
-    mc_now, _ = self.cfg.GetMasterCandidateStats()
-    master_candidate = mc_now < cp_size
+    if self.op.readd:
+      exceptions = [node]
+    else:
+      exceptions = []
+    mc_now, mc_max = self.cfg.GetMasterCandidateStats(exceptions)
+    # the new node will increase mc_max with one, so:
+    mc_max = min(mc_max + 1, cp_size)
+    self.master_candidate = mc_now < mc_max
 
 
-    self.new_node = objects.Node(name=node,
-                                 primary_ip=primary_ip,
-                                 secondary_ip=secondary_ip,
-                                 master_candidate=master_candidate,
-                                 offline=False, drained=False)
+    if self.op.readd:
+      self.new_node = self.cfg.GetNodeInfo(node)
+      assert self.new_node is not None, "Can't retrieve locked node %s" % node
+    else:
+      self.new_node = objects.Node(name=node,
+                                   primary_ip=primary_ip,
+                                   secondary_ip=secondary_ip,
+                                   master_candidate=self.master_candidate,
+                                   offline=False, drained=False)
 
   def Exec(self, feedback_fn):
     """Adds the new node to the cluster.
 
   def Exec(self, feedback_fn):
     """Adds the new node to the cluster.
@@ -2179,6 +2352,20 @@ class LUAddNode(LogicalUnit):
     new_node = self.new_node
     node = new_node.name
 
     new_node = self.new_node
     node = new_node.name
 
+    # for re-adds, reset the offline/drained/master-candidate flags;
+    # we need to reset here, otherwise offline would prevent RPC calls
+    # later in the procedure; this also means that if the re-add
+    # fails, we are left with a non-offlined, broken node
+    if self.op.readd:
+      new_node.drained = new_node.offline = False
+      self.LogInfo("Readding a node, the offline/drained flags were reset")
+      # if we demote the node, we do cleanup later in the procedure
+      new_node.master_candidate = self.master_candidate
+
+    # notify the user about any possible mc promotion
+    if new_node.master_candidate:
+      self.LogInfo("Node will be a master candidate")
+
     # check connectivity
     result = self.rpc.call_version([node])[node]
     result.Raise()
     # check connectivity
     result = self.rpc.call_version([node])[node]
     result.Raise()
@@ -2218,7 +2405,8 @@ class LUAddNode(LogicalUnit):
                                " new node: %s" % msg)
 
     # Add node to our /etc/hosts, and add key to known_hosts
                                " new node: %s" % msg)
 
     # Add node to our /etc/hosts, and add key to known_hosts
-    utils.AddHostToEtcHosts(new_node.name)
+    if self.cfg.GetClusterInfo().modify_etc_hosts:
+      utils.AddHostToEtcHosts(new_node.name)
 
     if new_node.secondary_ip != new_node.primary_ip:
       result = self.rpc.call_node_has_ip_address(new_node.name,
 
     if new_node.secondary_ip != new_node.primary_ip:
       result = self.rpc.call_node_has_ip_address(new_node.name,
@@ -2242,7 +2430,8 @@ class LUAddNode(LogicalUnit):
                                  " for remote verification" % verifier)
       if result[verifier].data['nodelist']:
         for failed in result[verifier].data['nodelist']:
                                  " for remote verification" % verifier)
       if result[verifier].data['nodelist']:
         for failed in result[verifier].data['nodelist']:
-          feedback_fn("ssh/hostname verification failed %s -> %s" %
+          feedback_fn("ssh/hostname verification failed"
+                      " (checking from %s): %s" %
                       (verifier, result[verifier].data['nodelist'][failed]))
         raise errors.OpExecError("ssh/hostname verification failed.")
 
                       (verifier, result[verifier].data['nodelist'][failed]))
         raise errors.OpExecError("ssh/hostname verification failed.")
 
@@ -2274,6 +2463,15 @@ class LUAddNode(LogicalUnit):
 
     if self.op.readd:
       self.context.ReaddNode(new_node)
 
     if self.op.readd:
       self.context.ReaddNode(new_node)
+      # make sure we redistribute the config
+      self.cfg.Update(new_node)
+      # and make sure the new node will not have old files around
+      if not new_node.master_candidate:
+        result = self.rpc.call_node_demote_from_mc(new_node.name)
+        msg = result.RemoteFailMsg()
+        if msg:
+          self.LogWarning("Node failed to demote itself from master"
+                          " candidate status: %s" % msg)
     else:
       self.context.AddNode(new_node)
 
     else:
       self.context.AddNode(new_node)
 
@@ -2329,12 +2527,16 @@ class LUSetNodeParams(LogicalUnit):
     """
     node = self.node = self.cfg.GetNodeInfo(self.op.node_name)
 
     """
     node = self.node = self.cfg.GetNodeInfo(self.op.node_name)
 
+    if (self.op.master_candidate is not None or
+        self.op.drained is not None or
+        self.op.offline is not None):
+      # we can't change the master's node flags
+      if self.op.node_name == self.cfg.GetMasterNode():
+        raise errors.OpPrereqError("The master role can be changed"
+                                   " only via masterfailover")
+
     if ((self.op.master_candidate == False or self.op.offline == True or
          self.op.drained == True) and node.master_candidate):
     if ((self.op.master_candidate == False or self.op.offline == True or
          self.op.drained == True) and node.master_candidate):
-      # we will demote the node from master_candidate
-      if self.op.node_name == self.cfg.GetMasterNode():
-        raise errors.OpPrereqError("The master node has to be a"
-                                   " master candidate, online and not drained")
       cp_size = self.cfg.GetClusterInfo().candidate_pool_size
       num_candidates, _ = self.cfg.GetMasterCandidateStats()
       if num_candidates <= cp_size:
       cp_size = self.cfg.GetClusterInfo().candidate_pool_size
       num_candidates, _ = self.cfg.GetMasterCandidateStats()
       if num_candidates <= cp_size:
@@ -2392,6 +2594,10 @@ class LUSetNodeParams(LogicalUnit):
           node.master_candidate = False
           changed_mc = True
           result.append(("master_candidate", "auto-demotion due to drain"))
           node.master_candidate = False
           changed_mc = True
           result.append(("master_candidate", "auto-demotion due to drain"))
+          rrc = self.rpc.call_node_demote_from_mc(node.name)
+          msg = rrc.RemoteFailMsg()
+          if msg:
+            self.LogWarning("Node failed to demote itself: %s" % msg)
         if node.offline:
           node.offline = False
           result.append(("offline", "clear offline status due to drain"))
         if node.offline:
           node.offline = False
           result.append(("offline", "clear offline status due to drain"))
@@ -2437,10 +2643,15 @@ class LUQueryClusterInfo(NoHooksLU):
       "master": cluster.master_node,
       "default_hypervisor": cluster.default_hypervisor,
       "enabled_hypervisors": cluster.enabled_hypervisors,
       "master": cluster.master_node,
       "default_hypervisor": cluster.default_hypervisor,
       "enabled_hypervisors": cluster.enabled_hypervisors,
-      "hvparams": dict([(hypervisor, cluster.hvparams[hypervisor])
-                        for hypervisor in cluster.enabled_hypervisors]),
+      "hvparams": dict([(hypervisor_name, cluster.hvparams[hypervisor_name])
+                        for hypervisor_name in cluster.enabled_hypervisors]),
       "beparams": cluster.beparams,
       "candidate_pool_size": cluster.candidate_pool_size,
       "beparams": cluster.beparams,
       "candidate_pool_size": cluster.candidate_pool_size,
+      "default_bridge": cluster.default_bridge,
+      "master_netdev": cluster.master_netdev,
+      "volume_group_name": cluster.volume_group_name,
+      "file_storage_dir": cluster.file_storage_dir,
+      "tags": list(cluster.GetTags()),
       }
 
     return result
       }
 
     return result
@@ -2512,19 +2723,24 @@ class LUActivateInstanceDisks(NoHooksLU):
     assert self.instance is not None, \
       "Cannot retrieve locked instance %s" % self.op.instance_name
     _CheckNodeOnline(self, self.instance.primary_node)
     assert self.instance is not None, \
       "Cannot retrieve locked instance %s" % self.op.instance_name
     _CheckNodeOnline(self, self.instance.primary_node)
+    if not hasattr(self.op, "ignore_size"):
+      self.op.ignore_size = False
 
   def Exec(self, feedback_fn):
     """Activate the disks.
 
     """
 
   def Exec(self, feedback_fn):
     """Activate the disks.
 
     """
-    disks_ok, disks_info = _AssembleInstanceDisks(self, self.instance)
+    disks_ok, disks_info = \
+              _AssembleInstanceDisks(self, self.instance,
+                                     ignore_size=self.op.ignore_size)
     if not disks_ok:
       raise errors.OpExecError("Cannot activate block devices")
 
     return disks_info
 
 
     if not disks_ok:
       raise errors.OpExecError("Cannot activate block devices")
 
     return disks_info
 
 
-def _AssembleInstanceDisks(lu, instance, ignore_secondaries=False):
+def _AssembleInstanceDisks(lu, instance, ignore_secondaries=False,
+                           ignore_size=False):
   """Prepare the block devices for an instance.
 
   This sets up the block devices on all nodes.
   """Prepare the block devices for an instance.
 
   This sets up the block devices on all nodes.
@@ -2536,6 +2752,10 @@ def _AssembleInstanceDisks(lu, instance, ignore_secondaries=False):
   @type ignore_secondaries: boolean
   @param ignore_secondaries: if true, errors on secondary nodes
       won't result in an error return from the function
   @type ignore_secondaries: boolean
   @param ignore_secondaries: if true, errors on secondary nodes
       won't result in an error return from the function
+  @type ignore_size: boolean
+  @param ignore_size: if true, the current known size of the disk
+      will not be used during the disk activation, useful for cases
+      when the size is wrong
   @return: False if the operation failed, otherwise a list of
       (host, instance_visible_name, node_visible_name)
       with the mapping from node devices to instance devices
   @return: False if the operation failed, otherwise a list of
       (host, instance_visible_name, node_visible_name)
       with the mapping from node devices to instance devices
@@ -2556,6 +2776,9 @@ def _AssembleInstanceDisks(lu, instance, ignore_secondaries=False):
   # 1st pass, assemble on all nodes in secondary mode
   for inst_disk in instance.disks:
     for node, node_disk in inst_disk.ComputeNodeTree(instance.primary_node):
   # 1st pass, assemble on all nodes in secondary mode
   for inst_disk in instance.disks:
     for node, node_disk in inst_disk.ComputeNodeTree(instance.primary_node):
+      if ignore_size:
+        node_disk = node_disk.Copy()
+        node_disk.UnsetSize()
       lu.cfg.SetDiskID(node_disk, node)
       result = lu.rpc.call_blockdev_assemble(node, node_disk, iname, False)
       msg = result.RemoteFailMsg()
       lu.cfg.SetDiskID(node_disk, node)
       result = lu.rpc.call_blockdev_assemble(node, node_disk, iname, False)
       msg = result.RemoteFailMsg()
@@ -2573,6 +2796,9 @@ def _AssembleInstanceDisks(lu, instance, ignore_secondaries=False):
     for node, node_disk in inst_disk.ComputeNodeTree(instance.primary_node):
       if node != instance.primary_node:
         continue
     for node, node_disk in inst_disk.ComputeNodeTree(instance.primary_node):
       if node != instance.primary_node:
         continue
+      if ignore_size:
+        node_disk = node_disk.Copy()
+        node_disk.UnsetSize()
       lu.cfg.SetDiskID(node_disk, node)
       result = lu.rpc.call_blockdev_assemble(node, node_disk, iname, True)
       msg = result.RemoteFailMsg()
       lu.cfg.SetDiskID(node_disk, node)
       result = lu.rpc.call_blockdev_assemble(node, node_disk, iname, True)
       msg = result.RemoteFailMsg()
@@ -2597,7 +2823,7 @@ def _StartInstanceDisks(lu, instance, force):
   """Start the disks of an instance.
 
   """
   """Start the disks of an instance.
 
   """
-  disks_ok, dummy = _AssembleInstanceDisks(lu, instance,
+  disks_ok, _ = _AssembleInstanceDisks(lu, instance,
                                            ignore_secondaries=force)
   if not disks_ok:
     _ShutdownInstanceDisks(lu, instance)
                                            ignore_secondaries=force)
   if not disks_ok:
     _ShutdownInstanceDisks(lu, instance)
@@ -2755,15 +2981,48 @@ class LUStartupInstance(LogicalUnit):
     assert self.instance is not None, \
       "Cannot retrieve locked instance %s" % self.op.instance_name
 
     assert self.instance is not None, \
       "Cannot retrieve locked instance %s" % self.op.instance_name
 
+    # extra beparams
+    self.beparams = getattr(self.op, "beparams", {})
+    if self.beparams:
+      if not isinstance(self.beparams, dict):
+        raise errors.OpPrereqError("Invalid beparams passed: %s, expected"
+                                   " dict" % (type(self.beparams), ))
+      # fill the beparams dict
+      utils.ForceDictType(self.beparams, constants.BES_PARAMETER_TYPES)
+      self.op.beparams = self.beparams
+
+    # extra hvparams
+    self.hvparams = getattr(self.op, "hvparams", {})
+    if self.hvparams:
+      if not isinstance(self.hvparams, dict):
+        raise errors.OpPrereqError("Invalid hvparams passed: %s, expected"
+                                   " dict" % (type(self.hvparams), ))
+
+      # check hypervisor parameter syntax (locally)
+      cluster = self.cfg.GetClusterInfo()
+      utils.ForceDictType(self.hvparams, constants.HVS_PARAMETER_TYPES)
+      filled_hvp = cluster.FillDict(cluster.hvparams[instance.hypervisor],
+                                    instance.hvparams)
+      filled_hvp.update(self.hvparams)
+      hv_type = hypervisor.GetHypervisor(instance.hypervisor)
+      hv_type.CheckParameterSyntax(filled_hvp)
+      _CheckHVParams(self, instance.all_nodes, instance.hypervisor, filled_hvp)
+      self.op.hvparams = self.hvparams
+
     _CheckNodeOnline(self, instance.primary_node)
 
     bep = self.cfg.GetClusterInfo().FillBE(instance)
     _CheckNodeOnline(self, instance.primary_node)
 
     bep = self.cfg.GetClusterInfo().FillBE(instance)
-    # check bridges existance
+    # check bridges existence
     _CheckInstanceBridgesExist(self, instance)
 
     _CheckInstanceBridgesExist(self, instance)
 
-    _CheckNodeFreeMemory(self, instance.primary_node,
-                         "starting instance %s" % instance.name,
-                         bep[constants.BE_MEMORY], instance.hypervisor)
+    remote_info = self.rpc.call_instance_info(instance.primary_node,
+                                              instance.name,
+                                              instance.hypervisor)
+    remote_info.Raise()
+    if not remote_info.data:
+      _CheckNodeFreeMemory(self, instance.primary_node,
+                           "starting instance %s" % instance.name,
+                           bep[constants.BE_MEMORY], instance.hypervisor)
 
   def Exec(self, feedback_fn):
     """Start the instance.
 
   def Exec(self, feedback_fn):
     """Start the instance.
@@ -2778,7 +3037,8 @@ class LUStartupInstance(LogicalUnit):
 
     _StartInstanceDisks(self, instance, force)
 
 
     _StartInstanceDisks(self, instance, force)
 
-    result = self.rpc.call_instance_start(node_current, instance)
+    result = self.rpc.call_instance_start(node_current, instance,
+                                          self.hvparams, self.beparams)
     msg = result.RemoteFailMsg()
     if msg:
       _ShutdownInstanceDisks(self, instance)
     msg = result.RemoteFailMsg()
     if msg:
       _ShutdownInstanceDisks(self, instance)
@@ -2830,7 +3090,7 @@ class LURebootInstance(LogicalUnit):
 
     _CheckNodeOnline(self, instance.primary_node)
 
 
     _CheckNodeOnline(self, instance.primary_node)
 
-    # check bridges existance
+    # check bridges existence
     _CheckInstanceBridgesExist(self, instance)
 
   def Exec(self, feedback_fn):
     _CheckInstanceBridgesExist(self, instance)
 
   def Exec(self, feedback_fn):
@@ -2860,7 +3120,7 @@ class LURebootInstance(LogicalUnit):
                                  " full reboot: %s" % msg)
       _ShutdownInstanceDisks(self, instance)
       _StartInstanceDisks(self, instance, ignore_secondaries)
                                  " full reboot: %s" % msg)
       _ShutdownInstanceDisks(self, instance)
       _StartInstanceDisks(self, instance, ignore_secondaries)
-      result = self.rpc.call_instance_start(node_current, instance)
+      result = self.rpc.call_instance_start(node_current, instance, None, None)
       msg = result.RemoteFailMsg()
       if msg:
         _ShutdownInstanceDisks(self, instance)
       msg = result.RemoteFailMsg()
       if msg:
         _ShutdownInstanceDisks(self, instance)
@@ -2960,7 +3220,8 @@ class LUReinstallInstance(LogicalUnit):
     remote_info = self.rpc.call_instance_info(instance.primary_node,
                                               instance.name,
                                               instance.hypervisor)
     remote_info = self.rpc.call_instance_info(instance.primary_node,
                                               instance.name,
                                               instance.hypervisor)
-    if remote_info.failed or remote_info.data:
+    remote_info.Raise()
+    if remote_info.data:
       raise errors.OpPrereqError("Instance '%s' is running on the node %s" %
                                  (self.op.instance_name,
                                   instance.primary_node))
       raise errors.OpPrereqError("Instance '%s' is running on the node %s" %
                                  (self.op.instance_name,
                                   instance.primary_node))
@@ -3341,14 +3602,25 @@ class LUQueryInstances(NoHooksLU):
             val = live_data[instance.name].get("memory", "?")
           else:
             val = "-"
             val = live_data[instance.name].get("memory", "?")
           else:
             val = "-"
+        elif field == "vcpus":
+          val = i_be[constants.BE_VCPUS]
         elif field == "disk_template":
           val = instance.disk_template
         elif field == "ip":
         elif field == "disk_template":
           val = instance.disk_template
         elif field == "ip":
-          val = instance.nics[0].ip
+          if instance.nics:
+            val = instance.nics[0].ip
+          else:
+            val = None
         elif field == "bridge":
         elif field == "bridge":
-          val = instance.nics[0].bridge
+          if instance.nics:
+            val = instance.nics[0].bridge
+          else:
+            val = None
         elif field == "mac":
         elif field == "mac":
-          val = instance.nics[0].mac
+          if instance.nics:
+            val = instance.nics[0].mac
+          else:
+            val = None
         elif field == "sda_size" or field == "sdb_size":
           idx = ord(field[2]) - ord('a')
           try:
         elif field == "sda_size" or field == "sdb_size":
           idx = ord(field[2]) - ord('a')
           try:
@@ -3415,9 +3687,10 @@ class LUQueryInstances(NoHooksLU):
                 else:
                   assert False, "Unhandled NIC parameter"
           else:
                 else:
                   assert False, "Unhandled NIC parameter"
           else:
-            assert False, "Unhandled variable parameter"
+            assert False, ("Declared but unhandled variable parameter '%s'" %
+                           field)
         else:
         else:
-          raise errors.ParameterError(field)
+          assert False, "Declared but unhandled parameter '%s'" % field
         iout.append(val)
       output.append(iout)
 
         iout.append(val)
       output.append(iout)
 
@@ -3478,12 +3751,17 @@ class LUFailoverInstance(LogicalUnit):
     target_node = secondary_nodes[0]
     _CheckNodeOnline(self, target_node)
     _CheckNodeNotDrained(self, target_node)
     target_node = secondary_nodes[0]
     _CheckNodeOnline(self, target_node)
     _CheckNodeNotDrained(self, target_node)
-    # check memory requirements on the secondary node
-    _CheckNodeFreeMemory(self, target_node, "failing over instance %s" %
-                         instance.name, bep[constants.BE_MEMORY],
-                         instance.hypervisor)
 
 
-    # check bridge existance
+    if instance.admin_up:
+      # check memory requirements on the secondary node
+      _CheckNodeFreeMemory(self, target_node, "failing over instance %s" %
+                           instance.name, bep[constants.BE_MEMORY],
+                           instance.hypervisor)
+    else:
+      self.LogInfo("Not checking memory on the secondary node as"
+                   " instance will not be started")
+
+    # check bridge existence
     brlist = [nic.bridge for nic in instance.nics]
     result = self.rpc.call_bridges_exist(target_node, brlist)
     result.Raise()
     brlist = [nic.bridge for nic in instance.nics]
     result = self.rpc.call_bridges_exist(target_node, brlist)
     result.Raise()
@@ -3543,14 +3821,14 @@ class LUFailoverInstance(LogicalUnit):
       logging.info("Starting instance %s on node %s",
                    instance.name, target_node)
 
       logging.info("Starting instance %s on node %s",
                    instance.name, target_node)
 
-      disks_ok, dummy = _AssembleInstanceDisks(self, instance,
+      disks_ok, _ = _AssembleInstanceDisks(self, instance,
                                                ignore_secondaries=True)
       if not disks_ok:
         _ShutdownInstanceDisks(self, instance)
         raise errors.OpExecError("Can't activate the instance's disks")
 
       feedback_fn("* starting the instance on the target node")
                                                ignore_secondaries=True)
       if not disks_ok:
         _ShutdownInstanceDisks(self, instance)
         raise errors.OpExecError("Can't activate the instance's disks")
 
       feedback_fn("* starting the instance on the target node")
-      result = self.rpc.call_instance_start(target_node, instance)
+      result = self.rpc.call_instance_start(target_node, instance, None, None)
       msg = result.RemoteFailMsg()
       if msg:
         _ShutdownInstanceDisks(self, instance)
       msg = result.RemoteFailMsg()
       if msg:
         _ShutdownInstanceDisks(self, instance)
@@ -3621,7 +3899,7 @@ class LUMigrateInstance(LogicalUnit):
                          instance.name, i_be[constants.BE_MEMORY],
                          instance.hypervisor)
 
                          instance.name, i_be[constants.BE_MEMORY],
                          instance.hypervisor)
 
-    # check bridge existance
+    # check bridge existence
     brlist = [nic.bridge for nic in instance.nics]
     result = self.rpc.call_bridges_exist(target_node, brlist)
     if result.failed or not result.data:
     brlist = [nic.bridge for nic in instance.nics]
     result = self.rpc.call_bridges_exist(target_node, brlist)
     if result.failed or not result.data:
@@ -4057,7 +4335,7 @@ def _GenerateDiskTemplate(lu, template_name,
     if len(secondary_nodes) != 0:
       raise errors.ProgrammerError("Wrong template configuration")
 
     if len(secondary_nodes) != 0:
       raise errors.ProgrammerError("Wrong template configuration")
 
-    names = _GenerateUniqueNames(lu, [".disk%d" % i
+    names = _GenerateUniqueNames(lu, [".disk%d" % (base_index + i)
                                       for i in range(disk_count)])
     for idx, disk in enumerate(disk_info):
       disk_index = idx + base_index
                                       for i in range(disk_count)])
     for idx, disk in enumerate(disk_info):
       disk_index = idx + base_index
@@ -4074,7 +4352,7 @@ def _GenerateDiskTemplate(lu, template_name,
       [primary_node, remote_node] * len(disk_info), instance_name)
 
     names = []
       [primary_node, remote_node] * len(disk_info), instance_name)
 
     names = []
-    for lv_prefix in _GenerateUniqueNames(lu, [".disk%d" % i
+    for lv_prefix in _GenerateUniqueNames(lu, [".disk%d" % (base_index + i)
                                                for i in range(disk_count)]):
       names.append(lv_prefix + "_data")
       names.append(lv_prefix + "_meta")
                                                for i in range(disk_count)]):
       names.append(lv_prefix + "_data")
       names.append(lv_prefix + "_meta")
@@ -4234,8 +4512,8 @@ def _CheckHVParams(lu, nodenames, hvname, hvparams):
       continue
     msg = info.RemoteFailMsg()
     if msg:
       continue
     msg = info.RemoteFailMsg()
     if msg:
-      raise errors.OpPrereqError("Hypervisor parameter validation failed:"
-                                 " %s" % msg)
+      raise errors.OpPrereqError("Hypervisor parameter validation"
+                                 " failed on node %s: %s" % (node, msg))
 
 
 class LUCreateInstance(LogicalUnit):
 
 
 class LUCreateInstance(LogicalUnit):
@@ -4300,6 +4578,7 @@ class LUCreateInstance(LogicalUnit):
                                   self.op.hvparams)
     hv_type = hypervisor.GetHypervisor(self.op.hypervisor)
     hv_type.CheckParameterSyntax(filled_hvp)
                                   self.op.hvparams)
     hv_type = hypervisor.GetHypervisor(self.op.hypervisor)
     hv_type.CheckParameterSyntax(filled_hvp)
+    self.hv_full = filled_hvp
 
     # fill and remember the beparams dict
     utils.ForceDictType(self.op.beparams, constants.BES_PARAMETER_TYPES)
 
     # fill and remember the beparams dict
     utils.ForceDictType(self.op.beparams, constants.BES_PARAMETER_TYPES)
@@ -4341,6 +4620,12 @@ class LUCreateInstance(LogicalUnit):
         if not utils.IsValidMac(mac.lower()):
           raise errors.OpPrereqError("Invalid MAC address specified: %s" %
                                      mac)
         if not utils.IsValidMac(mac.lower()):
           raise errors.OpPrereqError("Invalid MAC address specified: %s" %
                                      mac)
+        else:
+          # or validate/reserve the current one
+          if self.cfg.IsMacInUse(mac):
+            raise errors.OpPrereqError("MAC address %s already in use"
+                                       " in cluster" % mac)
+
       # bridge verification
       bridge = nic.get("bridge", None)
       if bridge is None:
       # bridge verification
       bridge = nic.get("bridge", None)
       if bridge is None:
@@ -4477,6 +4762,9 @@ class LUCreateInstance(LogicalUnit):
       nics=[(n.ip, n.bridge, n.mac) for n in self.nics],
       disk_template=self.op.disk_template,
       disks=[(d["size"], d["mode"]) for d in self.disks],
       nics=[(n.ip, n.bridge, n.mac) for n in self.nics],
       disk_template=self.op.disk_template,
       disks=[(d["size"], d["mode"]) for d in self.disks],
+      bep=self.be_full,
+      hvp=self.hv_full,
+      hypervisor_name=self.op.hypervisor,
     ))
 
     nl = ([self.cfg.GetMasterNode(), self.op.pnode] +
     ))
 
     nl = ([self.cfg.GetMasterNode(), self.op.pnode] +
@@ -4643,7 +4931,7 @@ class LUCreateInstance(LogicalUnit):
     # os verification
     result = self.rpc.call_os_get(pnode.name, self.op.os_type)
     result.Raise()
     # os verification
     result = self.rpc.call_os_get(pnode.name, self.op.os_type)
     result.Raise()
-    if not isinstance(result.data, objects.OS):
+    if not isinstance(result.data, objects.OS) or not result.data:
       raise errors.OpPrereqError("OS '%s' not in supported os list for"
                                  " primary node"  % self.op.os_type)
 
       raise errors.OpPrereqError("OS '%s' not in supported os list for"
                                  " primary node"  % self.op.os_type)
 
@@ -4794,7 +5082,7 @@ class LUCreateInstance(LogicalUnit):
       self.cfg.Update(iobj)
       logging.info("Starting instance %s on node %s", instance, pnode_name)
       feedback_fn("* starting instance...")
       self.cfg.Update(iobj)
       logging.info("Starting instance %s on node %s", instance, pnode_name)
       feedback_fn("* starting instance...")
-      result = self.rpc.call_instance_start(pnode_name, iobj)
+      result = self.rpc.call_instance_start(pnode_name, iobj, None, None)
       msg = result.RemoteFailMsg()
       if msg:
         raise errors.OpExecError("Could not start instance: %s" % msg)
       msg = result.RemoteFailMsg()
       if msg:
         raise errors.OpExecError("Could not start instance: %s" % msg)
@@ -5295,7 +5583,6 @@ class LUReplaceDisks(LogicalUnit):
     logging.debug("Allocated minors %s" % (minors,))
     self.proc.LogStep(4, steps_total, "changing drbd configuration")
     for idx, (dev, new_minor) in enumerate(zip(instance.disks, minors)):
     logging.debug("Allocated minors %s" % (minors,))
     self.proc.LogStep(4, steps_total, "changing drbd configuration")
     for idx, (dev, new_minor) in enumerate(zip(instance.disks, minors)):
-      size = dev.size
       info("activating a new drbd on %s for disk/%d" % (new_node, idx))
       # create new devices on new_node; note that we create two IDs:
       # one without port, so the drbd will be activated without
       info("activating a new drbd on %s for disk/%d" % (new_node, idx))
       # create new devices on new_node; note that we create two IDs:
       # one without port, so the drbd will be activated without
@@ -5315,7 +5602,8 @@ class LUReplaceDisks(LogicalUnit):
                     new_net_id)
       new_drbd = objects.Disk(dev_type=constants.LD_DRBD8,
                               logical_id=new_alone_id,
                     new_net_id)
       new_drbd = objects.Disk(dev_type=constants.LD_DRBD8,
                               logical_id=new_alone_id,
-                              children=dev.children)
+                              children=dev.children,
+                              size=dev.size)
       try:
         _CreateSingleBlockDev(self, new_node, instance, new_drbd,
                               _GetInstanceInfoText(instance), False)
       try:
         _CreateSingleBlockDev(self, new_node, instance, new_drbd,
                               _GetInstanceInfoText(instance), False)
@@ -5615,6 +5903,7 @@ class LUQueryInstanceData(NoHooksLU):
       "sstatus": dev_sstatus,
       "children": dev_children,
       "mode": dev.mode,
       "sstatus": dev_sstatus,
       "children": dev_children,
       "mode": dev.mode,
+      "size": dev.size,
       }
 
     return data
       }
 
     return data
@@ -5828,7 +6117,7 @@ class LUSetInstanceParams(LogicalUnit):
     This only checks the instance list against the existing names.
 
     """
     This only checks the instance list against the existing names.
 
     """
-    force = self.force = self.op.force
+    self.force = self.op.force
 
     # checking the new params on the primary/secondary nodes
 
 
     # checking the new params on the primary/secondary nodes
 
@@ -5898,7 +6187,7 @@ class LUSetInstanceParams(LogicalUnit):
         self.warn.append("Can't get info from primary node %s" % pnode)
       else:
         if not instance_info.failed and instance_info.data:
         self.warn.append("Can't get info from primary node %s" % pnode)
       else:
         if not instance_info.failed and instance_info.data:
-          current_mem = instance_info.data['memory']
+          current_mem = int(instance_info.data['memory'])
         else:
           # Assume instance not running
           # (there is a slight race condition here, but it's not very probable,
         else:
           # Assume instance not running
           # (there is a slight race condition here, but it's not very probable,
@@ -6151,7 +6440,7 @@ class LUExportInstance(LogicalUnit):
     # remove it from its current node. In the future we could fix this by:
     #  - making a tasklet to search (share-lock all), then create the new one,
     #    then one to remove, after
     # remove it from its current node. In the future we could fix this by:
     #  - making a tasklet to search (share-lock all), then create the new one,
     #    then one to remove, after
-    #  - removing the removal operation altoghether
+    #  - removing the removal operation altogether
     self.needed_locks[locking.LEVEL_NODE] = locking.ALL_SET
 
   def DeclareLocks(self, level):
     self.needed_locks[locking.LEVEL_NODE] = locking.ALL_SET
 
   def DeclareLocks(self, level):
@@ -6225,13 +6514,15 @@ class LUExportInstance(LogicalUnit):
     for disk in instance.disks:
       self.cfg.SetDiskID(disk, src_node)
 
     for disk in instance.disks:
       self.cfg.SetDiskID(disk, src_node)
 
+    # per-disk results
+    dresults = []
     try:
     try:
-      for disk in instance.disks:
+      for idx, disk in enumerate(instance.disks):
         # new_dev_name will be a snapshot of an lvm leaf of the one we passed
         new_dev_name = self.rpc.call_blockdev_snapshot(src_node, disk)
         if new_dev_name.failed or not new_dev_name.data:
         # new_dev_name will be a snapshot of an lvm leaf of the one we passed
         new_dev_name = self.rpc.call_blockdev_snapshot(src_node, disk)
         if new_dev_name.failed or not new_dev_name.data:
-          self.LogWarning("Could not snapshot block device %s on node %s",
-                          disk.logical_id[1], src_node)
+          self.LogWarning("Could not snapshot disk/%d on node %s",
+                          idx, src_node)
           snap_disks.append(False)
         else:
           new_dev = objects.Disk(dev_type=constants.LD_LV, size=disk.size,
           snap_disks.append(False)
         else:
           new_dev = objects.Disk(dev_type=constants.LD_LV, size=disk.size,
@@ -6242,7 +6533,7 @@ class LUExportInstance(LogicalUnit):
 
     finally:
       if self.op.shutdown and instance.admin_up:
 
     finally:
       if self.op.shutdown and instance.admin_up:
-        result = self.rpc.call_instance_start(src_node, instance)
+        result = self.rpc.call_instance_start(src_node, instance, None, None)
         msg = result.RemoteFailMsg()
         if msg:
           _ShutdownInstanceDisks(self, instance)
         msg = result.RemoteFailMsg()
         if msg:
           _ShutdownInstanceDisks(self, instance)
@@ -6256,18 +6547,24 @@ class LUExportInstance(LogicalUnit):
         result = self.rpc.call_snapshot_export(src_node, dev, dst_node.name,
                                                instance, cluster_name, idx)
         if result.failed or not result.data:
         result = self.rpc.call_snapshot_export(src_node, dev, dst_node.name,
                                                instance, cluster_name, idx)
         if result.failed or not result.data:
-          self.LogWarning("Could not export block device %s from node %s to"
-                          " node %s", dev.logical_id[1], src_node,
-                          dst_node.name)
+          self.LogWarning("Could not export disk/%d from node %s to"
+                          " node %s", idx, src_node, dst_node.name)
+          dresults.append(False)
+        else:
+          dresults.append(True)
         msg = self.rpc.call_blockdev_remove(src_node, dev).RemoteFailMsg()
         if msg:
         msg = self.rpc.call_blockdev_remove(src_node, dev).RemoteFailMsg()
         if msg:
-          self.LogWarning("Could not remove snapshot block device %s from node"
-                          " %s: %s", dev.logical_id[1], src_node, msg)
+          self.LogWarning("Could not remove snapshot for disk/%d from node"
+                          " %s: %s", idx, src_node, msg)
+      else:
+        dresults.append(False)
 
     result = self.rpc.call_finalize_export(dst_node.name, instance, snap_disks)
 
     result = self.rpc.call_finalize_export(dst_node.name, instance, snap_disks)
+    fin_resu = True
     if result.failed or not result.data:
       self.LogWarning("Could not finalize export for instance %s on node %s",
                       instance.name, dst_node.name)
     if result.failed or not result.data:
       self.LogWarning("Could not finalize export for instance %s on node %s",
                       instance.name, dst_node.name)
+      fin_resu = False
 
     nodelist = self.cfg.GetNodeList()
     nodelist.remove(dst_node.name)
 
     nodelist = self.cfg.GetNodeList()
     nodelist.remove(dst_node.name)
@@ -6284,6 +6581,7 @@ class LUExportInstance(LogicalUnit):
           if not self.rpc.call_export_remove(node, instance.name):
             self.LogWarning("Could not remove older export for instance %s"
                             " on node %s", instance.name, node)
           if not self.rpc.call_export_remove(node, instance.name):
             self.LogWarning("Could not remove older export for instance %s"
                             " on node %s", instance.name, node)
+    return fin_resu, dresults
 
 
 class LURemoveExport(NoHooksLU):
 
 
 class LURemoveExport(NoHooksLU):
@@ -6648,7 +6946,7 @@ class IAllocator(object):
         "master_candidate": ninfo.master_candidate,
         }
 
         "master_candidate": ninfo.master_candidate,
         }
 
-      if not ninfo.offline:
+      if not (ninfo.offline or ninfo.drained):
         nresult.Raise()
         if not isinstance(nresult.data, dict):
           raise errors.OpExecError("Can't get data for node %s" % nname)
         nresult.Raise()
         if not isinstance(nresult.data, dict):
           raise errors.OpExecError("Can't get data for node %s" % nname)
@@ -6711,6 +7009,8 @@ class IAllocator(object):
         "disk_template": iinfo.disk_template,
         "hypervisor": iinfo.hypervisor,
         }
         "disk_template": iinfo.disk_template,
         "hypervisor": iinfo.hypervisor,
         }
+      pir["disk_space_total"] = _ComputeDiskSize(iinfo.disk_template,
+                                                 pir["disks"])
       instance_data[iinfo.name] = pir
 
     data["instances"] = instance_data
       instance_data[iinfo.name] = pir
 
     data["instances"] = instance_data
@@ -6803,7 +7103,6 @@ class IAllocator(object):
     """
     if call_fn is None:
       call_fn = self.lu.rpc.call_iallocator_runner
     """
     if call_fn is None:
       call_fn = self.lu.rpc.call_iallocator_runner
-    data = self.in_text
 
     result = call_fn(self.lu.cfg.GetMasterNode(), name, self.in_text)
     result.Raise()
 
     result = call_fn(self.lu.cfg.GetMasterNode(), name, self.in_text)
     result.Raise()