Merge branch 'stable-2.1' into devel-2.1
[ganeti-local] / lib / cmdlib.py
index ddbc48d..cb922eb 100644 (file)
 
 """Module implementing the master-side code."""
 
-# pylint: disable-msg=W0613,W0201
+# pylint: disable-msg=W0201
+
+# W0201 since most LU attributes are defined in CheckPrereq or similar
+# functions
 
 import os
 import os.path
@@ -87,9 +90,9 @@ class LogicalUnit(object):
     self.recalculate_locks = {}
     self.__ssh = None
     # logging
-    self.LogWarning = processor.LogWarning
-    self.LogInfo = processor.LogInfo
-    self.LogStep = processor.LogStep
+    self.LogWarning = processor.LogWarning # pylint: disable-msg=C0103
+    self.LogInfo = processor.LogInfo # pylint: disable-msg=C0103
+    self.LogStep = processor.LogStep # pylint: disable-msg=C0103
     # support for dry-run
     self.dry_run_result = None
 
@@ -100,7 +103,7 @@ class LogicalUnit(object):
       attr_val = getattr(op, attr_name, None)
       if attr_val is None:
         raise errors.OpPrereqError("Required parameter '%s' missing" %
-                                   attr_name)
+                                   attr_name, errors.ECODE_INVAL)
 
     self.CheckArguments()
 
@@ -277,6 +280,9 @@ class LogicalUnit(object):
         and hook results
 
     """
+    # API must be kept, thus we ignore the unused argument and could
+    # be a function warnings
+    # pylint: disable-msg=W0613,R0201
     return lu_result
 
   def _ExpandAndLockInstance(self):
@@ -297,7 +303,7 @@ class LogicalUnit(object):
     expanded_name = self.cfg.ExpandInstanceName(self.op.instance_name)
     if expanded_name is None:
       raise errors.OpPrereqError("Instance '%s' not known" %
-                                  self.op.instance_name)
+                                 self.op.instance_name, errors.ECODE_NOENT)
     self.needed_locks[locking.LEVEL_INSTANCE] = expanded_name
     self.op.instance_name = expanded_name
 
@@ -347,7 +353,7 @@ class LogicalUnit(object):
     del self.recalculate_locks[locking.LEVEL_NODE]
 
 
-class NoHooksLU(LogicalUnit):
+class NoHooksLU(LogicalUnit): # pylint: disable-msg=W0223
   """Simple LU which runs no hooks.
 
   This LU is intended as a parent for other LogicalUnits which will
@@ -357,6 +363,14 @@ class NoHooksLU(LogicalUnit):
   HPATH = None
   HTYPE = None
 
+  def BuildHooksEnv(self):
+    """Empty BuildHooksEnv for NoHooksLu.
+
+    This just raises an error.
+
+    """
+    assert False, "BuildHooksEnv called for NoHooksLUs"
+
 
 class Tasklet:
   """Tasklet base class.
@@ -417,7 +431,8 @@ def _GetWantedNodes(lu, nodes):
 
   """
   if not isinstance(nodes, list):
-    raise errors.OpPrereqError("Invalid argument type 'nodes'")
+    raise errors.OpPrereqError("Invalid argument type 'nodes'",
+                               errors.ECODE_INVAL)
 
   if not nodes:
     raise errors.ProgrammerError("_GetWantedNodes should only be called with a"
@@ -427,7 +442,8 @@ def _GetWantedNodes(lu, nodes):
   for name in nodes:
     node = lu.cfg.ExpandNodeName(name)
     if node is None:
-      raise errors.OpPrereqError("No such node name '%s'" % name)
+      raise errors.OpPrereqError("No such node name '%s'" % name,
+                                 errors.ECODE_NOENT)
     wanted.append(node)
 
   return utils.NiceSort(wanted)
@@ -447,7 +463,8 @@ def _GetWantedInstances(lu, instances):
 
   """
   if not isinstance(instances, list):
-    raise errors.OpPrereqError("Invalid argument type 'instances'")
+    raise errors.OpPrereqError("Invalid argument type 'instances'",
+                               errors.ECODE_INVAL)
 
   if instances:
     wanted = []
@@ -455,7 +472,8 @@ def _GetWantedInstances(lu, instances):
     for name in instances:
       instance = lu.cfg.ExpandInstanceName(name)
       if instance is None:
-        raise errors.OpPrereqError("No such instance name '%s'" % name)
+        raise errors.OpPrereqError("No such instance name '%s'" % name,
+                                   errors.ECODE_NOENT)
       wanted.append(instance)
 
   else:
@@ -479,7 +497,7 @@ def _CheckOutputFields(static, dynamic, selected):
   delta = f.NonMatching(selected)
   if delta:
     raise errors.OpPrereqError("Unknown output fields selected: %s"
-                               % ",".join(delta))
+                               % ",".join(delta), errors.ECODE_INVAL)
 
 
 def _CheckBooleanOpField(op, name):
@@ -492,10 +510,25 @@ def _CheckBooleanOpField(op, name):
   val = getattr(op, name, None)
   if not (val is None or isinstance(val, bool)):
     raise errors.OpPrereqError("Invalid boolean parameter '%s' (%s)" %
-                               (name, str(val)))
+                               (name, str(val)), errors.ECODE_INVAL)
   setattr(op, name, val)
 
 
+def _CheckGlobalHvParams(params):
+  """Validates that given hypervisor params are not global ones.
+
+  This will ensure that instances don't get customised versions of
+  global params.
+
+  """
+  used_globals = constants.HVC_GLOBALS.intersection(params)
+  if used_globals:
+    msg = ("The following hypervisor parameters are global and cannot"
+           " be customized at instance level, please modify them at"
+           " cluster level: %s" % utils.CommaJoin(used_globals))
+    raise errors.OpPrereqError(msg, errors.ECODE_INVAL)
+
+
 def _CheckNodeOnline(lu, node):
   """Ensure that a given node is online.
 
@@ -505,7 +538,8 @@ def _CheckNodeOnline(lu, node):
 
   """
   if lu.cfg.GetNodeInfo(node).offline:
-    raise errors.OpPrereqError("Can't use offline node %s" % node)
+    raise errors.OpPrereqError("Can't use offline node %s" % node,
+                               errors.ECODE_INVAL)
 
 
 def _CheckNodeNotDrained(lu, node):
@@ -517,7 +551,8 @@ def _CheckNodeNotDrained(lu, node):
 
   """
   if lu.cfg.GetNodeInfo(node).drained:
-    raise errors.OpPrereqError("Can't use drained node %s" % node)
+    raise errors.OpPrereqError("Can't use drained node %s" % node,
+                               errors.ECODE_INVAL)
 
 
 def _BuildInstanceHookEnv(name, primary_node, secondary_nodes, os_type, status,
@@ -667,25 +702,36 @@ def _BuildInstanceHookEnvByObject(lu, instance, override=None):
   }
   if override:
     args.update(override)
-  return _BuildInstanceHookEnv(**args)
+  return _BuildInstanceHookEnv(**args) # pylint: disable-msg=W0142
 
 
-def _AdjustCandidatePool(lu):
+def _AdjustCandidatePool(lu, exceptions):
   """Adjust the candidate pool after node operations.
 
   """
-  mod_list = lu.cfg.MaintainCandidatePool()
+  mod_list = lu.cfg.MaintainCandidatePool(exceptions)
   if mod_list:
     lu.LogInfo("Promoted nodes to master candidate role: %s",
-               ", ".join(node.name for node in mod_list))
+               utils.CommaJoin(node.name for node in mod_list))
     for name in mod_list:
       lu.context.ReaddNode(name)
-  mc_now, mc_max = lu.cfg.GetMasterCandidateStats()
+  mc_now, mc_max, _ = lu.cfg.GetMasterCandidateStats(exceptions)
   if mc_now > mc_max:
     lu.LogInfo("Note: more nodes are candidates (%d) than desired (%d)" %
                (mc_now, mc_max))
 
 
+def _DecideSelfPromotion(lu, exceptions=None):
+  """Decide whether I should promote myself as a master candidate.
+
+  """
+  cp_size = lu.cfg.GetClusterInfo().candidate_pool_size
+  mc_now, mc_should, _ = lu.cfg.GetMasterCandidateStats(exceptions)
+  # the new node will increase mc_max with one, so:
+  mc_should = min(mc_should + 1, cp_size)
+  return mc_now < mc_should
+
+
 def _CheckNicsBridgesExist(lu, target_nics, target_node,
                                profile=constants.PP_DEFAULT):
   """Check that the brigdes needed by a list of nics exist.
@@ -699,7 +745,7 @@ def _CheckNicsBridgesExist(lu, target_nics, target_node,
   if brlist:
     result = lu.rpc.call_bridges_exist(target_node, brlist)
     result.Raise("Error checking bridges on destination node '%s'" %
-                 target_node, prereq=True)
+                 target_node, prereq=True, ecode=errors.ECODE_ENVIRON)
 
 
 def _CheckInstanceBridgesExist(lu, instance, node=None):
@@ -711,6 +757,27 @@ def _CheckInstanceBridgesExist(lu, instance, node=None):
   _CheckNicsBridgesExist(lu, instance.nics, node)
 
 
+def _CheckOSVariant(os_obj, name):
+  """Check whether an OS name conforms to the os variants specification.
+
+  @type os_obj: L{objects.OS}
+  @param os_obj: OS object to check
+  @type name: string
+  @param name: OS name passed by the user, to check for validity
+
+  """
+  if not os_obj.supported_variants:
+    return
+  try:
+    variant = name.split("+", 1)[1]
+  except IndexError:
+    raise errors.OpPrereqError("OS name must include a variant",
+                               errors.ECODE_INVAL)
+
+  if variant not in os_obj.supported_variants:
+    raise errors.OpPrereqError("Unsupported OS variant", errors.ECODE_INVAL)
+
+
 def _GetNodeInstancesInner(cfg, fn):
   return [i for i in cfg.GetAllInstancesInfo().values() if fn(i)]
 
@@ -759,7 +826,7 @@ def _FindFaultyInstanceDisks(cfg, rpc, instance, node_name, prereq):
 
   result = rpc.call_blockdev_getmirrorstatus(node_name, instance.disks)
   result.Raise("Failed to get disk status from node %s" % node_name,
-               prereq=prereq)
+               prereq=prereq, ecode=errors.ECODE_ENVIRON)
 
   for idx, bdev_status in enumerate(result.payload):
     if bdev_status and bdev_status.ldisk_status == constants.LDS_FAULTY:
@@ -825,30 +892,37 @@ class LUDestroyCluster(LogicalUnit):
     nodelist = self.cfg.GetNodeList()
     if len(nodelist) != 1 or nodelist[0] != master:
       raise errors.OpPrereqError("There are still %d node(s) in"
-                                 " this cluster." % (len(nodelist) - 1))
+                                 " this cluster." % (len(nodelist) - 1),
+                                 errors.ECODE_INVAL)
     instancelist = self.cfg.GetInstanceList()
     if instancelist:
       raise errors.OpPrereqError("There are still %d instance(s) in"
-                                 " this cluster." % len(instancelist))
+                                 " this cluster." % len(instancelist),
+                                 errors.ECODE_INVAL)
 
   def Exec(self, feedback_fn):
     """Destroys the cluster.
 
     """
     master = self.cfg.GetMasterNode()
+    modify_ssh_setup = self.cfg.GetClusterInfo().modify_ssh_setup
 
     # Run post hooks on master node before it's removed
     hm = self.proc.hmclass(self.rpc.call_hooks_runner, self)
     try:
       hm.RunPhase(constants.HOOKS_PHASE_POST, [master])
     except:
+      # pylint: disable-msg=W0702
       self.LogWarning("Errors occurred running hooks on %s" % master)
 
     result = self.rpc.call_node_stop_master(master, False)
     result.Raise("Could not disable the master role")
-    priv_key, pub_key, _ = ssh.GetUserFiles(constants.GANETI_RUNAS)
-    utils.CreateBackup(priv_key)
-    utils.CreateBackup(pub_key)
+
+    if modify_ssh_setup:
+      priv_key, pub_key, _ = ssh.GetUserFiles(constants.GANETI_RUNAS)
+      utils.CreateBackup(priv_key)
+      utils.CreateBackup(pub_key)
+
     return master
 
 
@@ -858,9 +932,39 @@ class LUVerifyCluster(LogicalUnit):
   """
   HPATH = "cluster-verify"
   HTYPE = constants.HTYPE_CLUSTER
-  _OP_REQP = ["skip_checks"]
+  _OP_REQP = ["skip_checks", "verbose", "error_codes", "debug_simulate_errors"]
   REQ_BGL = False
 
+  TCLUSTER = "cluster"
+  TNODE = "node"
+  TINSTANCE = "instance"
+
+  ECLUSTERCFG = (TCLUSTER, "ECLUSTERCFG")
+  EINSTANCEBADNODE = (TINSTANCE, "EINSTANCEBADNODE")
+  EINSTANCEDOWN = (TINSTANCE, "EINSTANCEDOWN")
+  EINSTANCELAYOUT = (TINSTANCE, "EINSTANCELAYOUT")
+  EINSTANCEMISSINGDISK = (TINSTANCE, "EINSTANCEMISSINGDISK")
+  EINSTANCEMISSINGDISK = (TINSTANCE, "EINSTANCEMISSINGDISK")
+  EINSTANCEWRONGNODE = (TINSTANCE, "EINSTANCEWRONGNODE")
+  ENODEDRBD = (TNODE, "ENODEDRBD")
+  ENODEFILECHECK = (TNODE, "ENODEFILECHECK")
+  ENODEHOOKS = (TNODE, "ENODEHOOKS")
+  ENODEHV = (TNODE, "ENODEHV")
+  ENODELVM = (TNODE, "ENODELVM")
+  ENODEN1 = (TNODE, "ENODEN1")
+  ENODENET = (TNODE, "ENODENET")
+  ENODEORPHANINSTANCE = (TNODE, "ENODEORPHANINSTANCE")
+  ENODEORPHANLV = (TNODE, "ENODEORPHANLV")
+  ENODERPC = (TNODE, "ENODERPC")
+  ENODESSH = (TNODE, "ENODESSH")
+  ENODEVERSION = (TNODE, "ENODEVERSION")
+  ENODESETUP = (TNODE, "ENODESETUP")
+  ENODETIME = (TNODE, "ENODETIME")
+
+  ETYPE_FIELD = "code"
+  ETYPE_ERROR = "ERROR"
+  ETYPE_WARNING = "WARNING"
+
   def ExpandNames(self):
     self.needed_locks = {
       locking.LEVEL_NODE: locking.ALL_SET,
@@ -868,9 +972,45 @@ class LUVerifyCluster(LogicalUnit):
     }
     self.share_locks = dict.fromkeys(locking.LEVELS, 1)
 
+  def _Error(self, ecode, item, msg, *args, **kwargs):
+    """Format an error message.
+
+    Based on the opcode's error_codes parameter, either format a
+    parseable error code, or a simpler error string.
+
+    This must be called only from Exec and functions called from Exec.
+
+    """
+    ltype = kwargs.get(self.ETYPE_FIELD, self.ETYPE_ERROR)
+    itype, etxt = ecode
+    # first complete the msg
+    if args:
+      msg = msg % args
+    # then format the whole message
+    if self.op.error_codes:
+      msg = "%s:%s:%s:%s:%s" % (ltype, etxt, itype, item, msg)
+    else:
+      if item:
+        item = " " + item
+      else:
+        item = ""
+      msg = "%s: %s%s: %s" % (ltype, itype, item, msg)
+    # and finally report it via the feedback_fn
+    self._feedback_fn("  - %s" % msg)
+
+  def _ErrorIf(self, cond, *args, **kwargs):
+    """Log an error message if the passed condition is True.
+
+    """
+    cond = bool(cond) or self.op.debug_simulate_errors
+    if cond:
+      self._Error(*args, **kwargs)
+    # do not mark the operation as failed for WARN cases only
+    if kwargs.get(self.ETYPE_FIELD, self.ETYPE_ERROR) == self.ETYPE_ERROR:
+      self.bad = self.bad or cond
+
   def _VerifyNode(self, nodeinfo, file_list, local_cksum,
-                  node_result, feedback_fn, master_files,
-                  drbd_map, vg_name):
+                  node_result, master_files, drbd_map, vg_name):
     """Run multiple tests against a node.
 
     Test list:
@@ -885,7 +1025,6 @@ class LUVerifyCluster(LogicalUnit):
     @param file_list: required list of files
     @param local_cksum: dictionary of local files and their checksums
     @param node_result: the results from the node
-    @param feedback_fn: function used to accumulate results
     @param master_files: list of files that only masters should have
     @param drbd_map: the useddrbd minors for this node, in
         form of minor: (instance, must_exist) which correspond to instances
@@ -894,138 +1033,154 @@ class LUVerifyCluster(LogicalUnit):
 
     """
     node = nodeinfo.name
+    _ErrorIf = self._ErrorIf # pylint: disable-msg=C0103
 
     # main result, node_result should be a non-empty dict
-    if not node_result or not isinstance(node_result, dict):
-      feedback_fn("  - ERROR: unable to verify node %s." % (node,))
-      return True
+    test = not node_result or not isinstance(node_result, dict)
+    _ErrorIf(test, self.ENODERPC, node,
+                  "unable to verify node: no data returned")
+    if test:
+      return
 
     # compares ganeti version
     local_version = constants.PROTOCOL_VERSION
     remote_version = node_result.get('version', None)
-    if not (remote_version and isinstance(remote_version, (list, tuple)) and
-            len(remote_version) == 2):
-      feedback_fn("  - ERROR: connection to %s failed" % (node))
-      return True
+    test = not (remote_version and
+                isinstance(remote_version, (list, tuple)) and
+                len(remote_version) == 2)
+    _ErrorIf(test, self.ENODERPC, node,
+             "connection to node returned invalid data")
+    if test:
+      return
 
-    if local_version != remote_version[0]:
-      feedback_fn("  - ERROR: incompatible protocol versions: master %s,"
-                  " node %s %s" % (local_version, node, remote_version[0]))
-      return True
+    test = local_version != remote_version[0]
+    _ErrorIf(test, self.ENODEVERSION, node,
+             "incompatible protocol versions: master %s,"
+             " node %s", local_version, remote_version[0])
+    if test:
+      return
 
     # node seems compatible, we can actually try to look into its results
 
-    bad = False
-
     # full package version
-    if constants.RELEASE_VERSION != remote_version[1]:
-      feedback_fn("  - WARNING: software version mismatch: master %s,"
-                  " node %s %s" %
-                  (constants.RELEASE_VERSION, node, remote_version[1]))
+    self._ErrorIf(constants.RELEASE_VERSION != remote_version[1],
+                  self.ENODEVERSION, node,
+                  "software version mismatch: master %s, node %s",
+                  constants.RELEASE_VERSION, remote_version[1],
+                  code=self.ETYPE_WARNING)
 
     # checks vg existence and size > 20G
     if vg_name is not None:
       vglist = node_result.get(constants.NV_VGLIST, None)
-      if not vglist:
-        feedback_fn("  - ERROR: unable to check volume groups on node %s." %
-                        (node,))
-        bad = True
-      else:
+      test = not vglist
+      _ErrorIf(test, self.ENODELVM, node, "unable to check volume groups")
+      if not test:
         vgstatus = utils.CheckVolumeGroupSize(vglist, vg_name,
                                               constants.MIN_VG_SIZE)
-        if vgstatus:
-          feedback_fn("  - ERROR: %s on node %s" % (vgstatus, node))
-          bad = True
+        _ErrorIf(vgstatus, self.ENODELVM, node, vgstatus)
 
     # checks config file checksum
 
     remote_cksum = node_result.get(constants.NV_FILELIST, None)
-    if not isinstance(remote_cksum, dict):
-      bad = True
-      feedback_fn("  - ERROR: node hasn't returned file checksum data")
-    else:
+    test = not isinstance(remote_cksum, dict)
+    _ErrorIf(test, self.ENODEFILECHECK, node,
+             "node hasn't returned file checksum data")
+    if not test:
       for file_name in file_list:
         node_is_mc = nodeinfo.master_candidate
-        must_have_file = file_name not in master_files
-        if file_name not in remote_cksum:
-          if node_is_mc or must_have_file:
-            bad = True
-            feedback_fn("  - ERROR: file '%s' missing" % file_name)
-        elif remote_cksum[file_name] != local_cksum[file_name]:
-          if node_is_mc or must_have_file:
-            bad = True
-            feedback_fn("  - ERROR: file '%s' has wrong checksum" % file_name)
-          else:
-            # not candidate and this is not a must-have file
-            bad = True
-            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:
-            feedback_fn("  - ERROR: file '%s' should not exist on non master"
-                        " candidates" % file_name)
+        must_have = (file_name not in master_files) or node_is_mc
+        # missing
+        test1 = file_name not in remote_cksum
+        # invalid checksum
+        test2 = not test1 and remote_cksum[file_name] != local_cksum[file_name]
+        # existing and good
+        test3 = not test1 and remote_cksum[file_name] == local_cksum[file_name]
+        _ErrorIf(test1 and must_have, self.ENODEFILECHECK, node,
+                 "file '%s' missing", file_name)
+        _ErrorIf(test2 and must_have, self.ENODEFILECHECK, node,
+                 "file '%s' has wrong checksum", file_name)
+        # not candidate and this is not a must-have file
+        _ErrorIf(test2 and not must_have, self.ENODEFILECHECK, node,
+                 "file '%s' should not exist on non master"
+                 " candidates (and the file is outdated)", file_name)
+        # all good, except non-master/non-must have combination
+        _ErrorIf(test3 and not must_have, self.ENODEFILECHECK, node,
+                 "file '%s' should not exist"
+                 " on non master candidates", file_name)
 
     # checks ssh to any
 
-    if constants.NV_NODELIST not in node_result:
-      bad = True
-      feedback_fn("  - ERROR: node hasn't returned node ssh connectivity data")
-    else:
+    test = constants.NV_NODELIST not in node_result
+    _ErrorIf(test, self.ENODESSH, node,
+             "node hasn't returned node ssh connectivity data")
+    if not test:
       if node_result[constants.NV_NODELIST]:
-        bad = True
-        for node in node_result[constants.NV_NODELIST]:
-          feedback_fn("  - ERROR: ssh communication with node '%s': %s" %
-                          (node, node_result[constants.NV_NODELIST][node]))
-
-    if constants.NV_NODENETTEST not in node_result:
-      bad = True
-      feedback_fn("  - ERROR: node hasn't returned node tcp connectivity data")
-    else:
+        for a_node, a_msg in node_result[constants.NV_NODELIST].items():
+          _ErrorIf(True, self.ENODESSH, node,
+                   "ssh communication with node '%s': %s", a_node, a_msg)
+
+    test = constants.NV_NODENETTEST not in node_result
+    _ErrorIf(test, self.ENODENET, node,
+             "node hasn't returned node tcp connectivity data")
+    if not test:
       if node_result[constants.NV_NODENETTEST]:
-        bad = True
         nlist = utils.NiceSort(node_result[constants.NV_NODENETTEST].keys())
-        for node in nlist:
-          feedback_fn("  - ERROR: tcp communication with node '%s': %s" %
-                          (node, node_result[constants.NV_NODENETTEST][node]))
+        for anode in nlist:
+          _ErrorIf(True, self.ENODENET, node,
+                   "tcp communication with node '%s': %s",
+                   anode, node_result[constants.NV_NODENETTEST][anode])
 
     hyp_result = node_result.get(constants.NV_HYPERVISOR, None)
     if isinstance(hyp_result, dict):
       for hv_name, hv_result in hyp_result.iteritems():
-        if hv_result is not None:
-          feedback_fn("  - ERROR: hypervisor %s verify failure: '%s'" %
-                      (hv_name, hv_result))
+        test = hv_result is not None
+        _ErrorIf(test, self.ENODEHV, node,
+                 "hypervisor %s verify failure: '%s'", hv_name, hv_result)
 
     # check used drbd list
     if vg_name is not None:
       used_minors = node_result.get(constants.NV_DRBDLIST, [])
-      if not isinstance(used_minors, (tuple, list)):
-        feedback_fn("  - ERROR: cannot parse drbd status file: %s" %
-                    str(used_minors))
-      else:
+      test = not isinstance(used_minors, (tuple, list))
+      _ErrorIf(test, self.ENODEDRBD, node,
+               "cannot parse drbd status file: %s", str(used_minors))
+      if not test:
         for minor, (iname, must_exist) in drbd_map.items():
-          if minor not in used_minors and must_exist:
-            feedback_fn("  - ERROR: drbd minor %d of instance %s is"
-                        " not active" % (minor, iname))
-            bad = True
+          test = minor not in used_minors and must_exist
+          _ErrorIf(test, self.ENODEDRBD, node,
+                   "drbd minor %d of instance %s is not active",
+                   minor, iname)
         for minor in used_minors:
-          if minor not in drbd_map:
-            feedback_fn("  - ERROR: unallocated drbd minor %d is in use" %
-                        minor)
-            bad = True
-
-    return bad
+          test = minor not in drbd_map
+          _ErrorIf(test, self.ENODEDRBD, node,
+                   "unallocated drbd minor %d is in use", minor)
+    test = node_result.get(constants.NV_NODESETUP,
+                           ["Missing NODESETUP results"])
+    _ErrorIf(test, self.ENODESETUP, node, "node setup error: %s",
+             "; ".join(test))
+
+    # check pv names
+    if vg_name is not None:
+      pvlist = node_result.get(constants.NV_PVLIST, None)
+      test = pvlist is None
+      _ErrorIf(test, self.ENODELVM, node, "Can't get PV list from node")
+      if not test:
+        # check that ':' is not present in PV names, since it's a
+        # special character for lvcreate (denotes the range of PEs to
+        # use on the PV)
+        for _, pvname, owner_vg in pvlist:
+          test = ":" in pvname
+          _ErrorIf(test, self.ENODELVM, node, "Invalid character ':' in PV"
+                   " '%s' of VG '%s'", pvname, owner_vg)
 
   def _VerifyInstance(self, instance, instanceconfig, node_vol_is,
-                      node_instance, feedback_fn, n_offline):
+                      node_instance, n_offline):
     """Verify an instance.
 
     This function checks to see if the required block devices are
     available on the instance's node.
 
     """
-    bad = False
-
+    _ErrorIf = self._ErrorIf # pylint: disable-msg=C0103
     node_current = instanceconfig.primary_node
 
     node_vol_should = {}
@@ -1036,69 +1191,57 @@ class LUVerifyCluster(LogicalUnit):
         # ignore missing volumes on offline nodes
         continue
       for volume in node_vol_should[node]:
-        if node not in node_vol_is or volume not in node_vol_is[node]:
-          feedback_fn("  - ERROR: volume %s missing on node %s" %
-                          (volume, node))
-          bad = True
+        test = node not in node_vol_is or volume not in node_vol_is[node]
+        _ErrorIf(test, self.EINSTANCEMISSINGDISK, instance,
+                 "volume %s missing on node %s", volume, node)
 
     if instanceconfig.admin_up:
-      if ((node_current not in node_instance or
-          not instance in node_instance[node_current]) and
-          node_current not in n_offline):
-        feedback_fn("  - ERROR: instance %s not running on node %s" %
-                        (instance, node_current))
-        bad = True
+      test = ((node_current not in node_instance or
+               not instance in node_instance[node_current]) and
+              node_current not in n_offline)
+      _ErrorIf(test, self.EINSTANCEDOWN, instance,
+               "instance not running on its primary node %s",
+               node_current)
 
     for node in node_instance:
       if (not node == node_current):
-        if instance in node_instance[node]:
-          feedback_fn("  - ERROR: instance %s should not run on node %s" %
-                          (instance, node))
-          bad = True
+        test = instance in node_instance[node]
+        _ErrorIf(test, self.EINSTANCEWRONGNODE, instance,
+                 "instance should not run on node %s", node)
 
-    return bad
-
-  def _VerifyOrphanVolumes(self, node_vol_should, node_vol_is, feedback_fn):
+  def _VerifyOrphanVolumes(self, node_vol_should, node_vol_is):
     """Verify if there are any unknown volumes in the cluster.
 
     The .os, .swap and backup volumes are ignored. All other volumes are
     reported as unknown.
 
     """
-    bad = False
-
     for node in node_vol_is:
       for volume in node_vol_is[node]:
-        if node not in node_vol_should or volume not in node_vol_should[node]:
-          feedback_fn("  - ERROR: volume %s on node %s should not exist" %
-                      (volume, node))
-          bad = True
-    return bad
+        test = (node not in node_vol_should or
+                volume not in node_vol_should[node])
+        self._ErrorIf(test, self.ENODEORPHANLV, node,
+                      "volume %s is unknown", volume)
 
-  def _VerifyOrphanInstances(self, instancelist, node_instance, feedback_fn):
+  def _VerifyOrphanInstances(self, instancelist, node_instance):
     """Verify the list of running instances.
 
     This checks what instances are running but unknown to the cluster.
 
     """
-    bad = False
     for node in node_instance:
-      for runninginstance in node_instance[node]:
-        if runninginstance not in instancelist:
-          feedback_fn("  - ERROR: instance %s on node %s should not exist" %
-                          (runninginstance, node))
-          bad = True
-    return bad
-
-  def _VerifyNPlusOneMemory(self, node_info, instance_cfg, feedback_fn):
+      for o_inst in node_instance[node]:
+        test = o_inst not in instancelist
+        self._ErrorIf(test, self.ENODEORPHANINSTANCE, node,
+                      "instance %s on node %s should not exist", o_inst, node)
+
+  def _VerifyNPlusOneMemory(self, node_info, instance_cfg):
     """Verify N+1 Memory Resilience.
 
     Check that if one single node dies we can still start all the instances it
     was primary for.
 
     """
-    bad = False
-
     for node, nodeinfo in node_info.iteritems():
       # This code checks that every node which is now listed as secondary has
       # enough memory to host all instances it is supposed to should a single
@@ -1114,11 +1257,10 @@ class LUVerifyCluster(LogicalUnit):
           bep = self.cfg.GetClusterInfo().FillBE(instance_cfg[instance])
           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 accommodate"
-                      " failovers should node %s fail" % (node, prinode))
-          bad = True
-    return bad
+        test = nodeinfo['mfree'] < needed_mem
+        self._ErrorIf(test, self.ENODEN1, node,
+                      "not enough memory on to accommodate"
+                      " failovers should peer node %s fail", prinode)
 
   def CheckPrereq(self):
     """Check prerequisites.
@@ -1129,7 +1271,8 @@ class LUVerifyCluster(LogicalUnit):
     """
     self.skip_set = frozenset(self.op.skip_checks)
     if not constants.VERIFY_OPTIONAL_CHECKS.issuperset(self.skip_set):
-      raise errors.OpPrereqError("Invalid checks to be skipped specified")
+      raise errors.OpPrereqError("Invalid checks to be skipped specified",
+                                 errors.ECODE_INVAL)
 
   def BuildHooksEnv(self):
     """Build hooks env.
@@ -1151,10 +1294,13 @@ class LUVerifyCluster(LogicalUnit):
     """Verify integrity of cluster, performing various test on nodes.
 
     """
-    bad = False
+    self.bad = False
+    _ErrorIf = self._ErrorIf # pylint: disable-msg=C0103
+    verbose = self.op.verbose
+    self._feedback_fn = feedback_fn
     feedback_fn("* Verifying global settings")
     for msg in self.cfg.VerifyConfig():
-      feedback_fn("  - ERROR: %s" % msg)
+      _ErrorIf(True, self.ECLUSTERCFG, None, msg)
 
     vg_name = self.cfg.GetVGName()
     hypervisors = self.cfg.GetClusterInfo().enabled_hypervisors
@@ -1195,23 +1341,36 @@ class LUVerifyCluster(LogicalUnit):
       constants.NV_INSTANCELIST: hypervisors,
       constants.NV_VERSION: None,
       constants.NV_HVINFO: self.cfg.GetHypervisorType(),
+      constants.NV_NODESETUP: None,
+      constants.NV_TIME: None,
       }
+
     if vg_name is not None:
       node_verify_param[constants.NV_VGLIST] = None
       node_verify_param[constants.NV_LVLIST] = vg_name
+      node_verify_param[constants.NV_PVLIST] = [vg_name]
       node_verify_param[constants.NV_DRBDLIST] = None
+
+    # Due to the way our RPC system works, exact response times cannot be
+    # guaranteed (e.g. a broken node could run into a timeout). By keeping the
+    # time before and after executing the request, we can at least have a time
+    # window.
+    nvinfo_starttime = time.time()
     all_nvinfo = self.rpc.call_node_verify(nodelist, node_verify_param,
                                            self.cfg.GetClusterName())
+    nvinfo_endtime = time.time()
 
     cluster = self.cfg.GetClusterInfo()
     master_node = self.cfg.GetMasterNode()
     all_drbd_map = self.cfg.ComputeDRBDMap()
 
+    feedback_fn("* Verifying node status")
     for node_i in nodeinfo:
       node = node_i.name
 
       if node_i.offline:
-        feedback_fn("* Skipping offline node %s" % (node,))
+        if verbose:
+          feedback_fn("* Skipping offline node %s" % (node,))
         n_offline.append(node)
         continue
 
@@ -1224,62 +1383,81 @@ class LUVerifyCluster(LogicalUnit):
         n_drained.append(node)
       else:
         ntype = "regular"
-      feedback_fn("* Verifying node %s (%s)" % (node, ntype))
+      if verbose:
+        feedback_fn("* Verifying node %s (%s)" % (node, ntype))
 
       msg = all_nvinfo[node].fail_msg
+      _ErrorIf(msg, self.ENODERPC, node, "while contacting node: %s", msg)
       if msg:
-        feedback_fn("  - ERROR: while contacting node %s: %s" % (node, msg))
-        bad = True
         continue
 
       nresult = all_nvinfo[node].payload
       node_drbd = {}
       for minor, instance in all_drbd_map[node].items():
-        if instance not in instanceinfo:
-          feedback_fn("  - ERROR: ghost instance '%s' in temporary DRBD map" %
-                      instance)
+        test = instance not in instanceinfo
+        _ErrorIf(test, self.ECLUSTERCFG, None,
+                 "ghost instance '%s' in temporary DRBD map", instance)
           # ghost instance should not be running, but otherwise we
           # don't give double warnings (both ghost instance and
           # unallocated minor in use)
+        if test:
           node_drbd[minor] = (instance, False)
         else:
           instance = instanceinfo[instance]
           node_drbd[minor] = (instance.name, instance.admin_up)
-      result = self._VerifyNode(node_i, file_names, local_checksums,
-                                nresult, feedback_fn, master_files,
-                                node_drbd, vg_name)
-      bad = bad or result
+
+      self._VerifyNode(node_i, file_names, local_checksums,
+                       nresult, master_files, node_drbd, vg_name)
 
       lvdata = nresult.get(constants.NV_LVLIST, "Missing LV data")
       if vg_name is None:
         node_volume[node] = {}
       elif isinstance(lvdata, basestring):
-        feedback_fn("  - ERROR: LVM problem on node %s: %s" %
-                    (node, utils.SafeEncode(lvdata)))
-        bad = True
+        _ErrorIf(True, self.ENODELVM, node, "LVM problem on node: %s",
+                 utils.SafeEncode(lvdata))
         node_volume[node] = {}
       elif not isinstance(lvdata, dict):
-        feedback_fn("  - ERROR: connection to %s failed (lvlist)" % (node,))
-        bad = True
+        _ErrorIf(True, self.ENODELVM, node, "rpc call to node failed (lvlist)")
         continue
       else:
         node_volume[node] = lvdata
 
       # node_instance
       idata = nresult.get(constants.NV_INSTANCELIST, None)
-      if not isinstance(idata, list):
-        feedback_fn("  - ERROR: connection to %s failed (instancelist)" %
-                    (node,))
-        bad = True
+      test = not isinstance(idata, list)
+      _ErrorIf(test, self.ENODEHV, node,
+               "rpc call to node failed (instancelist)")
+      if test:
         continue
 
       node_instance[node] = idata
 
       # node_info
       nodeinfo = nresult.get(constants.NV_HVINFO, None)
-      if not isinstance(nodeinfo, dict):
-        feedback_fn("  - ERROR: connection to %s failed (hvinfo)" % (node,))
-        bad = True
+      test = not isinstance(nodeinfo, dict)
+      _ErrorIf(test, self.ENODEHV, node, "rpc call to node failed (hvinfo)")
+      if test:
+        continue
+
+      # Node time
+      ntime = nresult.get(constants.NV_TIME, None)
+      try:
+        ntime_merged = utils.MergeTime(ntime)
+      except (ValueError, TypeError):
+        _ErrorIf(test, self.ENODETIME, node, "Node returned invalid time")
+
+      if ntime_merged < (nvinfo_starttime - constants.NODE_MAX_CLOCK_SKEW):
+        ntime_diff = abs(nvinfo_starttime - ntime_merged)
+      elif ntime_merged > (nvinfo_endtime + constants.NODE_MAX_CLOCK_SKEW):
+        ntime_diff = abs(ntime_merged - nvinfo_endtime)
+      else:
+        ntime_diff = None
+
+      _ErrorIf(ntime_diff is not None, self.ENODETIME, node,
+               "Node time diverges by at least %0.1fs from master node time",
+               ntime_diff)
+
+      if ntime_diff is not None:
         continue
 
       try:
@@ -1297,28 +1475,28 @@ class LUVerifyCluster(LogicalUnit):
         }
         # FIXME: devise a free space model for file based instances as well
         if vg_name is not None:
-          if (constants.NV_VGLIST not in nresult or
-              vg_name not in nresult[constants.NV_VGLIST]):
-            feedback_fn("  - ERROR: node %s didn't return data for the"
-                        " volume group '%s' - it is either missing or broken" %
-                        (node, vg_name))
-            bad = True
+          test = (constants.NV_VGLIST not in nresult or
+                  vg_name not in nresult[constants.NV_VGLIST])
+          _ErrorIf(test, self.ENODELVM, node,
+                   "node didn't return data for the volume group '%s'"
+                   " - it is either missing or broken", vg_name)
+          if test:
             continue
           node_info[node]["dfree"] = int(nresult[constants.NV_VGLIST][vg_name])
       except (ValueError, KeyError):
-        feedback_fn("  - ERROR: invalid nodeinfo value returned"
-                    " from node %s" % (node,))
-        bad = True
+        _ErrorIf(True, self.ENODERPC, node,
+                 "node returned invalid nodeinfo, check lvm/hypervisor")
         continue
 
     node_vol_should = {}
 
+    feedback_fn("* Verifying instance status")
     for instance in instancelist:
-      feedback_fn("* Verifying instance %s" % instance)
+      if verbose:
+        feedback_fn("* Verifying instance %s" % instance)
       inst_config = instanceinfo[instance]
-      result =  self._VerifyInstance(instance, inst_config, node_volume,
-                                     node_instance, feedback_fn, n_offline)
-      bad = bad or result
+      self._VerifyInstance(instance, inst_config, node_volume,
+                           node_instance, n_offline)
       inst_nodes_offline = []
 
       inst_config.MapLVsByNode(node_vol_should)
@@ -1326,12 +1504,11 @@ class LUVerifyCluster(LogicalUnit):
       instance_cfg[instance] = inst_config
 
       pnode = inst_config.primary_node
+      _ErrorIf(pnode not in node_info and pnode not in n_offline,
+               self.ENODERPC, pnode, "instance %s, connection to"
+               " primary node failed", instance)
       if pnode in node_info:
         node_info[pnode]['pinst'].append(instance)
-      elif pnode not in n_offline:
-        feedback_fn("  - ERROR: instance %s, connection to primary node"
-                    " %s failed" % (instance, pnode))
-        bad = True
 
       if pnode in n_offline:
         inst_nodes_offline.append(pnode)
@@ -1343,46 +1520,42 @@ class LUVerifyCluster(LogicalUnit):
       # FIXME: does not support file-backed instances
       if len(inst_config.secondary_nodes) == 0:
         i_non_redundant.append(instance)
-      elif len(inst_config.secondary_nodes) > 1:
-        feedback_fn("  - WARNING: multiple secondaries for instance %s"
-                    % instance)
+      _ErrorIf(len(inst_config.secondary_nodes) > 1,
+               self.EINSTANCELAYOUT, instance,
+               "instance has multiple secondary nodes", code="WARNING")
 
       if not cluster.FillBE(inst_config)[constants.BE_AUTO_BALANCE]:
         i_non_a_balanced.append(instance)
 
       for snode in inst_config.secondary_nodes:
+        _ErrorIf(snode not in node_info and snode not in n_offline,
+                 self.ENODERPC, snode,
+                 "instance %s, connection to secondary node"
+                 "failed", instance)
+
         if snode in node_info:
           node_info[snode]['sinst'].append(instance)
           if pnode not in node_info[snode]['sinst-by-pnode']:
             node_info[snode]['sinst-by-pnode'][pnode] = []
           node_info[snode]['sinst-by-pnode'][pnode].append(instance)
-        elif snode not in n_offline:
-          feedback_fn("  - ERROR: instance %s, connection to secondary node"
-                      " %s failed" % (instance, snode))
-          bad = True
+
         if snode in n_offline:
           inst_nodes_offline.append(snode)
 
-      if inst_nodes_offline:
-        # warn that the instance lives on offline nodes, and set bad=True
-        feedback_fn("  - ERROR: instance lives on offline node(s) %s" %
-                    ", ".join(inst_nodes_offline))
-        bad = True
+      # warn that the instance lives on offline nodes
+      _ErrorIf(inst_nodes_offline, self.EINSTANCEBADNODE, instance,
+               "instance lives on offline node(s) %s",
+               utils.CommaJoin(inst_nodes_offline))
 
     feedback_fn("* Verifying orphan volumes")
-    result = self._VerifyOrphanVolumes(node_vol_should, node_volume,
-                                       feedback_fn)
-    bad = bad or result
+    self._VerifyOrphanVolumes(node_vol_should, node_volume)
 
     feedback_fn("* Verifying remaining instances")
-    result = self._VerifyOrphanInstances(instancelist, node_instance,
-                                         feedback_fn)
-    bad = bad or result
+    self._VerifyOrphanInstances(instancelist, node_instance)
 
     if constants.VERIFY_NPLUSONE_MEM not in self.skip_set:
       feedback_fn("* Verifying N+1 Memory redundancy")
-      result = self._VerifyNPlusOneMemory(node_info, instance_cfg, feedback_fn)
-      bad = bad or result
+      self._VerifyNPlusOneMemory(node_info, instance_cfg)
 
     feedback_fn("* Other Notes")
     if i_non_redundant:
@@ -1399,7 +1572,7 @@ class LUVerifyCluster(LogicalUnit):
     if n_drained:
       feedback_fn("  - NOTICE: %d drained node(s) found." % len(n_drained))
 
-    return not bad
+    return not self.bad
 
   def HooksCallBack(self, phase, hooks_results, feedback_fn, lu_result):
     """Analyze the post-hooks' result
@@ -1422,33 +1595,28 @@ class LUVerifyCluster(LogicalUnit):
       # Used to change hooks' output to proper indentation
       indent_re = re.compile('^', re.M)
       feedback_fn("* Hooks Results")
-      if not hooks_results:
-        feedback_fn("  - ERROR: general communication failure")
-        lu_result = 1
-      else:
-        for node_name in hooks_results:
-          show_node_header = True
-          res = hooks_results[node_name]
-          msg = res.fail_msg
-          if msg:
-            if res.offline:
-              # no need to warn or set fail return value
-              continue
-            feedback_fn("    Communication failure in hooks execution: %s" %
-                        msg)
+      assert hooks_results, "invalid result from hooks"
+
+      for node_name in hooks_results:
+        res = hooks_results[node_name]
+        msg = res.fail_msg
+        test = msg and not res.offline
+        self._ErrorIf(test, self.ENODEHOOKS, node_name,
+                      "Communication failure in hooks execution: %s", msg)
+        if res.offline or msg:
+          # No need to investigate payload if node is offline or gave an error.
+          # override manually lu_result here as _ErrorIf only
+          # overrides self.bad
+          lu_result = 1
+          continue
+        for script, hkr, output in res.payload:
+          test = hkr == constants.HKR_FAIL
+          self._ErrorIf(test, self.ENODEHOOKS, node_name,
+                        "Script %s failed, output:", script)
+          if test:
+            output = indent_re.sub('      ', output)
+            feedback_fn("%s" % output)
             lu_result = 1
-            continue
-          for script, hkr, output in res.payload:
-            if hkr == constants.HKR_FAIL:
-              # The node header is only shown once, if there are
-              # failing hooks on that node
-              if show_node_header:
-                feedback_fn("  Node %s:" % node_name)
-                show_node_header = False
-              feedback_fn("    ERROR: Script %s failed, output:" % script)
-              output = indent_re.sub('      ', output)
-              feedback_fn("%s" % output)
-              lu_result = 1
 
       return lu_result
 
@@ -1520,7 +1688,7 @@ class LUVerifyDisks(NoHooksLU):
         continue
 
       lvs = node_res.payload
-      for lv_name, (_, lv_inactive, lv_online) in lvs.items():
+      for lv_name, (_, _, lv_online) in lvs.items():
         inst = nv_dict.pop((node, lv_name), None)
         if (not lv_online and inst is not None
             and inst.name not in res_instances):
@@ -1545,16 +1713,17 @@ class LURepairDiskSizes(NoHooksLU):
 
   def ExpandNames(self):
     if not isinstance(self.op.instances, list):
-      raise errors.OpPrereqError("Invalid argument type 'instances'")
+      raise errors.OpPrereqError("Invalid argument type 'instances'",
+                                 errors.ECODE_INVAL)
 
     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)
+          raise errors.OpPrereqError("Instance '%s' not known" % name,
+                                     errors.ECODE_NOENT)
         self.wanted_names.append(full_name)
-      self.needed_locks[locking.LEVEL_INSTANCE] = self.wanted_names
       self.needed_locks = {
         locking.LEVEL_NODE: [],
         locking.LEVEL_INSTANCE: self.wanted_names,
@@ -1584,6 +1753,29 @@ class LURepairDiskSizes(NoHooksLU):
     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.
 
@@ -1600,8 +1792,11 @@ class LURepairDiskSizes(NoHooksLU):
 
     changed = []
     for node, dskl in per_node_disks.items():
-      result = self.rpc.call_blockdev_getsizes(node, [v[2] for v in dskl])
-      if result.failed or result.fail_msg:
+      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.fail_msg:
         self.LogWarning("Failure in blockdev_getsizes call to node"
                         " %s, ignoring", node)
         continue
@@ -1624,8 +1819,11 @@ class LURepairDiskSizes(NoHooksLU):
                        " correcting: recorded %d, actual %d", idx,
                        instance.name, disk.size, size)
           disk.size = size
-          self.cfg.Update(instance)
+          self.cfg.Update(instance, feedback_fn)
           changed.append((instance.name, idx, size))
+        if self._EnsureChildSizes(disk):
+          self.cfg.Update(instance, feedback_fn)
+          changed.append((instance.name, idx, disk.size))
     return changed
 
 
@@ -1646,13 +1844,14 @@ class LURenameCluster(LogicalUnit):
       "NEW_NAME": self.op.name,
       }
     mn = self.cfg.GetMasterNode()
-    return env, [mn], [mn]
+    all_nodes = self.cfg.GetNodeList()
+    return env, [mn], all_nodes
 
   def CheckPrereq(self):
     """Verify that the passed name is a valid one.
 
     """
-    hostname = utils.HostInfo(self.op.name)
+    hostname = utils.GetHostInfo(self.op.name)
 
     new_name = hostname.name
     self.ip = new_ip = hostname.ip
@@ -1660,12 +1859,13 @@ class LURenameCluster(LogicalUnit):
     old_ip = self.cfg.GetMasterIP()
     if new_name == old_name and new_ip == old_ip:
       raise errors.OpPrereqError("Neither the name nor the IP address of the"
-                                 " cluster has changed")
+                                 " cluster has changed",
+                                 errors.ECODE_INVAL)
     if new_ip != old_ip:
       if utils.TcpPing(new_ip, constants.DEFAULT_NODED_PORT):
         raise errors.OpPrereqError("The given cluster IP address (%s) is"
                                    " reachable on the network. Aborting." %
-                                   new_ip)
+                                   new_ip, errors.ECODE_NOTUNIQUE)
 
     self.op.name = new_name
 
@@ -1685,7 +1885,7 @@ class LURenameCluster(LogicalUnit):
       cluster = self.cfg.GetClusterInfo()
       cluster.cluster_name = clustername
       cluster.master_ip = ip
-      self.cfg.Update(cluster)
+      self.cfg.Update(cluster, feedback_fn)
 
       # update the known hosts file
       ssh.WriteKnownHostsFile(self.cfg, constants.SSH_KNOWN_HOSTS_FILE)
@@ -1747,9 +1947,10 @@ class LUSetClusterParams(LogicalUnit):
         self.op.candidate_pool_size = int(self.op.candidate_pool_size)
       except (ValueError, TypeError), err:
         raise errors.OpPrereqError("Invalid candidate_pool_size value: %s" %
-                                   str(err))
+                                   str(err), errors.ECODE_INVAL)
       if self.op.candidate_pool_size < 1:
-        raise errors.OpPrereqError("At least one master candidate needed")
+        raise errors.OpPrereqError("At least one master candidate needed",
+                                   errors.ECODE_INVAL)
 
   def ExpandNames(self):
     # FIXME: in the future maybe other cluster params won't require checking on
@@ -1783,7 +1984,8 @@ class LUSetClusterParams(LogicalUnit):
         for disk in inst.disks:
           if _RecursiveCheckIfLVMBased(disk):
             raise errors.OpPrereqError("Cannot disable lvm storage while"
-                                       " lvm-based instances exist")
+                                       " lvm-based instances exist",
+                                       errors.ECODE_INVAL)
 
     node_list = self.acquired_locks[locking.LEVEL_NODE]
 
@@ -1802,7 +2004,7 @@ class LUSetClusterParams(LogicalUnit):
                                               constants.MIN_VG_SIZE)
         if vgstatus:
           raise errors.OpPrereqError("Error on node '%s': %s" %
-                                     (node, vgstatus))
+                                     (node, vgstatus), errors.ECODE_ENVIRON)
 
     self.cluster = cluster = self.cfg.GetClusterInfo()
     # validate params changes
@@ -1816,12 +2018,36 @@ class LUSetClusterParams(LogicalUnit):
       self.new_nicparams = objects.FillDict(
         cluster.nicparams[constants.PP_DEFAULT], self.op.nicparams)
       objects.NIC.CheckParameterSyntax(self.new_nicparams)
+      nic_errors = []
+
+      # check all instances for consistency
+      for instance in self.cfg.GetAllInstancesInfo().values():
+        for nic_idx, nic in enumerate(instance.nics):
+          params_copy = copy.deepcopy(nic.nicparams)
+          params_filled = objects.FillDict(self.new_nicparams, params_copy)
+
+          # check parameter syntax
+          try:
+            objects.NIC.CheckParameterSyntax(params_filled)
+          except errors.ConfigurationError, err:
+            nic_errors.append("Instance %s, nic/%d: %s" %
+                              (instance.name, nic_idx, err))
+
+          # if we're moving instances to routed, check that they have an ip
+          target_mode = params_filled[constants.NIC_MODE]
+          if target_mode == constants.NIC_MODE_ROUTED and not nic.ip:
+            nic_errors.append("Instance %s, nic/%d: routed nick with no ip" %
+                              (instance.name, nic_idx))
+      if nic_errors:
+        raise errors.OpPrereqError("Cannot apply the change, errors:\n%s" %
+                                   "\n".join(nic_errors))
 
     # hypervisor list/parameters
     self.new_hvparams = objects.FillDict(cluster.hvparams, {})
     if self.op.hvparams:
       if not isinstance(self.op.hvparams, dict):
-        raise errors.OpPrereqError("Invalid 'hvparams' parameter on input")
+        raise errors.OpPrereqError("Invalid 'hvparams' parameter on input",
+                                   errors.ECODE_INVAL)
       for hv_name, hv_dict in self.op.hvparams.items():
         if hv_name not in self.new_hvparams:
           self.new_hvparams[hv_name] = hv_dict
@@ -1832,12 +2058,14 @@ class LUSetClusterParams(LogicalUnit):
       self.hv_list = self.op.enabled_hypervisors
       if not self.hv_list:
         raise errors.OpPrereqError("Enabled hypervisors list must contain at"
-                                   " least one member")
+                                   " least one member",
+                                   errors.ECODE_INVAL)
       invalid_hvs = set(self.hv_list) - constants.HYPER_TYPES
       if invalid_hvs:
         raise errors.OpPrereqError("Enabled hypervisors contains invalid"
                                    " entries: %s" %
-                                   utils.CommaJoin(invalid_hvs))
+                                   utils.CommaJoin(invalid_hvs),
+                                   errors.ECODE_INVAL)
     else:
       self.hv_list = cluster.enabled_hypervisors
 
@@ -1878,9 +2106,9 @@ class LUSetClusterParams(LogicalUnit):
     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)
+      _AdjustCandidatePool(self, [])
 
-    self.cfg.Update(self.cluster)
+    self.cfg.Update(self.cluster, feedback_fn)
 
 
 def _RedistributeAncillaryFiles(lu, additional_nodes=None):
@@ -1901,6 +2129,7 @@ def _RedistributeAncillaryFiles(lu, additional_nodes=None):
     dist_nodes.extend(additional_nodes)
   if myself.name in dist_nodes:
     dist_nodes.remove(myself.name)
+
   # 2. Gather files to distribute
   dist_files = set([constants.ETC_HOSTS,
                     constants.SSH_KNOWN_HOSTS_FILE,
@@ -1950,11 +2179,11 @@ class LURedistributeConfig(NoHooksLU):
     """Redistribute the configuration.
 
     """
-    self.cfg.Update(self.cfg.GetClusterInfo())
+    self.cfg.Update(self.cfg.GetClusterInfo(), feedback_fn)
     _RedistributeAncillaryFiles(self)
 
 
-def _WaitForSync(lu, instance, oneshot=False, unlock=False):
+def _WaitForSync(lu, instance, oneshot=False):
   """Sleep and poll for an instance's disk to sync.
 
   """
@@ -1969,6 +2198,8 @@ def _WaitForSync(lu, instance, oneshot=False, unlock=False):
   for dev in instance.disks:
     lu.cfg.SetDiskID(dev, node)
 
+  # TODO: Convert to utils.Retry
+
   retries = 0
   degr_retries = 10 # in seconds, as we sleep 1 second each time
   while True:
@@ -2066,11 +2297,14 @@ class LUDiagnoseOS(NoHooksLU):
   _OP_REQP = ["output_fields", "names"]
   REQ_BGL = False
   _FIELDS_STATIC = utils.FieldSet()
-  _FIELDS_DYNAMIC = utils.FieldSet("name", "valid", "node_status")
+  _FIELDS_DYNAMIC = utils.FieldSet("name", "valid", "node_status", "variants")
+  # Fields that need calculation of global os validity
+  _FIELDS_NEEDVALID = frozenset(["valid", "variants"])
 
   def ExpandNames(self):
     if self.op.names:
-      raise errors.OpPrereqError("Selective OS query not supported")
+      raise errors.OpPrereqError("Selective OS query not supported",
+                                 errors.ECODE_INVAL)
 
     _CheckOutputFields(static=self._FIELDS_STATIC,
                        dynamic=self._FIELDS_DYNAMIC,
@@ -2089,10 +2323,9 @@ class LUDiagnoseOS(NoHooksLU):
     """
 
   @staticmethod
-  def _DiagnoseByOS(node_list, rlist):
+  def _DiagnoseByOS(rlist):
     """Remaps a per-node return list into an a per-os per-node dictionary
 
-    @param node_list: a list with the names of all nodes
     @param rlist: a map with node names as keys and OS objects as values
 
     @rtype: dict
@@ -2114,14 +2347,14 @@ class LUDiagnoseOS(NoHooksLU):
     for node_name, nr in rlist.items():
       if nr.fail_msg or not nr.payload:
         continue
-      for name, path, status, diagnose in nr.payload:
+      for name, path, status, diagnose, variants in nr.payload:
         if name not in all_os:
           # build a list of nodes for this os containing empty lists
           # for each node in node_list
           all_os[name] = {}
           for nname in good_nodes:
             all_os[name][nname] = []
-        all_os[name][node_name].append((path, status, diagnose))
+        all_os[name][node_name].append((path, status, diagnose, variants))
     return all_os
 
   def Exec(self, feedback_fn):
@@ -2130,20 +2363,40 @@ class LUDiagnoseOS(NoHooksLU):
     """
     valid_nodes = [node for node in self.cfg.GetOnlineNodeList()]
     node_data = self.rpc.call_os_diagnose(valid_nodes)
-    pol = self._DiagnoseByOS(valid_nodes, node_data)
+    pol = self._DiagnoseByOS(node_data)
     output = []
+    calc_valid = self._FIELDS_NEEDVALID.intersection(self.op.output_fields)
+    calc_variants = "variants" in self.op.output_fields
+
     for os_name, os_data in pol.items():
       row = []
+      if calc_valid:
+        valid = True
+        variants = None
+        for osl in os_data.values():
+          valid = valid and osl and osl[0][1]
+          if not valid:
+            variants = None
+            break
+          if calc_variants:
+            node_variants = osl[0][3]
+            if variants is None:
+              variants = node_variants
+            else:
+              variants = [v for v in variants if v in node_variants]
+
       for field in self.op.output_fields:
         if field == "name":
           val = os_name
         elif field == "valid":
-          val = utils.all([osl and osl[0][1] for osl in os_data.values()])
+          val = valid
         elif field == "node_status":
           # this is just a copy of the dict
           val = {}
           for node_name, nos_list in os_data.items():
             val[node_name] = nos_list
+        elif field == "variants":
+          val =  variants
         else:
           raise errors.ParameterError(field)
         row.append(val)
@@ -2172,8 +2425,11 @@ class LURemoveNode(LogicalUnit):
       "NODE_NAME": self.op.node_name,
       }
     all_nodes = self.cfg.GetNodeList()
-    if self.op.node_name in all_nodes:
+    try:
       all_nodes.remove(self.op.node_name)
+    except ValueError:
+      logging.warning("Node %s which is about to be removed not found"
+                      " in the all nodes list", self.op.node_name)
     return env, all_nodes, all_nodes
 
   def CheckPrereq(self):
@@ -2189,20 +2445,23 @@ class LURemoveNode(LogicalUnit):
     """
     node = self.cfg.GetNodeInfo(self.cfg.ExpandNodeName(self.op.node_name))
     if node is None:
-      raise errors.OpPrereqError, ("Node '%s' is unknown." % self.op.node_name)
+      raise errors.OpPrereqError("Node '%s' is unknown." % self.op.node_name,
+                                 errors.ECODE_NOENT)
 
     instance_list = self.cfg.GetInstanceList()
 
     masternode = self.cfg.GetMasterNode()
     if node.name == masternode:
       raise errors.OpPrereqError("Node is the master node,"
-                                 " you need to failover first.")
+                                 " you need to failover first.",
+                                 errors.ECODE_INVAL)
 
     for instance_name in instance_list:
       instance = self.cfg.GetInstanceInfo(instance_name)
       if node.name in instance.all_nodes:
         raise errors.OpPrereqError("Instance %s is still running on the node,"
-                                   " please remove first." % instance_name)
+                                   " please remove first." % instance_name,
+                                   errors.ECODE_INVAL)
     self.op.node_name = node.name
     self.node = node
 
@@ -2214,31 +2473,38 @@ class LURemoveNode(LogicalUnit):
     logging.info("Stopping the node daemon and removing configs from node %s",
                  node.name)
 
+    modify_ssh_setup = self.cfg.GetClusterInfo().modify_ssh_setup
+
+    # Promote nodes to master candidate as needed
+    _AdjustCandidatePool(self, exceptions=[node.name])
     self.context.RemoveNode(node.name)
 
     # Run post hooks on the node before it's removed
     hm = self.proc.hmclass(self.rpc.call_hooks_runner, self)
     try:
-      h_results = hm.RunPhase(constants.HOOKS_PHASE_POST, [node.name])
+      hm.RunPhase(constants.HOOKS_PHASE_POST, [node.name])
     except:
+      # pylint: disable-msg=W0702
       self.LogWarning("Errors occurred running hooks on %s" % node.name)
 
-    result = self.rpc.call_node_leave_cluster(node.name)
+    result = self.rpc.call_node_leave_cluster(node.name, modify_ssh_setup)
     msg = result.fail_msg
     if msg:
       self.LogWarning("Errors encountered on the remote node while leaving"
                       " the cluster: %s", msg)
 
-    # Promote nodes to master candidate as needed
-    _AdjustCandidatePool(self)
-
 
 class LUQueryNodes(NoHooksLU):
   """Logical unit for querying nodes.
 
   """
+  # pylint: disable-msg=W0142
   _OP_REQP = ["output_fields", "names", "use_locking"]
   REQ_BGL = False
+
+  _SIMPLE_FIELDS = ["name", "serial_no", "ctime", "mtime", "uuid",
+                    "master_candidate", "offline", "drained"]
+
   _FIELDS_DYNAMIC = utils.FieldSet(
     "dtotal", "dfree",
     "mtotal", "mnode", "mfree",
@@ -2246,16 +2512,12 @@ class LUQueryNodes(NoHooksLU):
     "ctotal", "cnodes", "csockets",
     )
 
-  _FIELDS_STATIC = utils.FieldSet(
-    "name", "pinst_cnt", "sinst_cnt",
+  _FIELDS_STATIC = utils.FieldSet(*[
+    "pinst_cnt", "sinst_cnt",
     "pinst_list", "sinst_list",
     "pip", "sip", "tags",
-    "serial_no", "ctime", "mtime",
-    "master_candidate",
     "master",
-    "offline",
-    "drained",
-    "role",
+    "role"] + _SIMPLE_FIELDS
     )
 
   def ExpandNames(self):
@@ -2277,7 +2539,6 @@ class LUQueryNodes(NoHooksLU):
       # if we don't request only static fields, we need to lock the nodes
       self.needed_locks[locking.LEVEL_NODE] = self.wanted
 
-
   def CheckPrereq(self):
     """Check prerequisites.
 
@@ -2338,10 +2599,9 @@ class LUQueryNodes(NoHooksLU):
     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 inst in inst_data.values():
         if inst.primary_node in node_to_primary:
           node_to_primary[inst.primary_node].add(inst.name)
         for secnode in inst.secondary_nodes:
@@ -2356,8 +2616,8 @@ class LUQueryNodes(NoHooksLU):
     for node in nodelist:
       node_output = []
       for field in self.op.output_fields:
-        if field == "name":
-          val = node.name
+        if field in self._SIMPLE_FIELDS:
+          val = getattr(node, field)
         elif field == "pinst_list":
           val = list(node_to_primary[node.name])
         elif field == "sinst_list":
@@ -2372,20 +2632,8 @@ class LUQueryNodes(NoHooksLU):
           val = node.secondary_ip
         elif field == "tags":
           val = list(node.GetTags())
-        elif field == "serial_no":
-          val = node.serial_no
-        elif field == "ctime":
-          val = node.ctime
-        elif field == "mtime":
-          val = node.mtime
-        elif field == "master_candidate":
-          val = node.master_candidate
         elif field == "master":
           val = node.name == master_node
-        elif field == "offline":
-          val = node.offline
-        elif field == "drained":
-          val = node.drained
         elif self._FIELDS_DYNAMIC.Matches(field):
           val = live_data[node.name].get(field, None)
         elif field == "role":
@@ -2499,18 +2747,17 @@ class LUQueryNodeStorage(NoHooksLU):
   """
   _OP_REQP = ["nodes", "storage_type", "output_fields"]
   REQ_BGL = False
-  _FIELDS_STATIC = utils.FieldSet("node")
+  _FIELDS_STATIC = utils.FieldSet(constants.SF_NODE)
 
   def ExpandNames(self):
     storage_type = self.op.storage_type
 
-    if storage_type not in constants.VALID_STORAGE_FIELDS:
-      raise errors.OpPrereqError("Unknown storage type: %s" % storage_type)
-
-    dynamic_fields = constants.VALID_STORAGE_FIELDS[storage_type]
+    if storage_type not in constants.VALID_STORAGE_TYPES:
+      raise errors.OpPrereqError("Unknown storage type: %s" % storage_type,
+                                 errors.ECODE_INVAL)
 
     _CheckOutputFields(static=self._FIELDS_STATIC,
-                       dynamic=utils.FieldSet(*dynamic_fields),
+                       dynamic=utils.FieldSet(*constants.VALID_STORAGE_FIELDS),
                        selected=self.op.output_fields)
 
     self.needed_locks = {}
@@ -2542,9 +2789,10 @@ class LUQueryNodeStorage(NoHooksLU):
     else:
       fields = [constants.SF_NAME] + self.op.output_fields
 
-    # Never ask for node as it's only known to the LU
-    while "node" in fields:
-      fields.remove("node")
+    # Never ask for node or type as it's only known to the LU
+    for extra in [constants.SF_NODE, constants.SF_TYPE]:
+      while extra in fields:
+        fields.remove(extra)
 
     field_idx = dict([(name, idx) for (idx, name) in enumerate(fields)])
     name_idx = field_idx[constants.SF_NAME]
@@ -2574,8 +2822,10 @@ class LUQueryNodeStorage(NoHooksLU):
         out = []
 
         for field in self.op.output_fields:
-          if field == "node":
+          if field == constants.SF_NODE:
             val = node
+          elif field == constants.SF_TYPE:
+            val = self.op.storage_type
           elif field in field_idx:
             val = row[field_idx[field]]
           else:
@@ -2598,13 +2848,15 @@ class LUModifyNodeStorage(NoHooksLU):
   def CheckArguments(self):
     node_name = self.cfg.ExpandNodeName(self.op.node_name)
     if node_name is None:
-      raise errors.OpPrereqError("Invalid node name '%s'" % self.op.node_name)
+      raise errors.OpPrereqError("Invalid node name '%s'" % self.op.node_name,
+                                 errors.ECODE_NOENT)
 
     self.op.node_name = node_name
 
     storage_type = self.op.storage_type
-    if storage_type not in constants.VALID_STORAGE_FIELDS:
-      raise errors.OpPrereqError("Unknown storage type: %s" % storage_type)
+    if storage_type not in constants.VALID_STORAGE_TYPES:
+      raise errors.OpPrereqError("Unknown storage type: %s" % storage_type,
+                                 errors.ECODE_INVAL)
 
   def ExpandNames(self):
     self.needed_locks = {
@@ -2621,13 +2873,15 @@ class LUModifyNodeStorage(NoHooksLU):
       modifiable = constants.MODIFIABLE_STORAGE_FIELDS[storage_type]
     except KeyError:
       raise errors.OpPrereqError("Storage units of type '%s' can not be"
-                                 " modified" % storage_type)
+                                 " modified" % storage_type,
+                                 errors.ECODE_INVAL)
 
     diff = set(self.op.changes.keys()) - modifiable
     if diff:
       raise errors.OpPrereqError("The following fields can not be modified for"
                                  " storage units of type '%s': %r" %
-                                 (storage_type, list(diff)))
+                                 (storage_type, list(diff)),
+                                 errors.ECODE_INVAL)
 
   def Exec(self, feedback_fn):
     """Computes the list of nodes and their attributes.
@@ -2679,7 +2933,7 @@ class LUAddNode(LogicalUnit):
     node_name = self.op.node_name
     cfg = self.cfg
 
-    dns_data = utils.HostInfo(node_name)
+    dns_data = utils.GetHostInfo(node_name)
 
     node = dns_data.name
     primary_ip = self.op.primary_ip = dns_data.ip
@@ -2687,15 +2941,17 @@ class LUAddNode(LogicalUnit):
     if secondary_ip is None:
       secondary_ip = primary_ip
     if not utils.IsValidIP(secondary_ip):
-      raise errors.OpPrereqError("Invalid secondary IP given")
+      raise errors.OpPrereqError("Invalid secondary IP given",
+                                 errors.ECODE_INVAL)
     self.op.secondary_ip = secondary_ip
 
     node_list = cfg.GetNodeList()
     if not self.op.readd and node in node_list:
       raise errors.OpPrereqError("Node %s is already in the configuration" %
-                                 node)
+                                 node, errors.ECODE_EXISTS)
     elif self.op.readd and node not in node_list:
-      raise errors.OpPrereqError("Node %s is not in the configuration" % node)
+      raise errors.OpPrereqError("Node %s is not in the configuration" % node,
+                                 errors.ECODE_NOENT)
 
     for existing_node_name in node_list:
       existing_node = cfg.GetNodeInfo(existing_node_name)
@@ -2704,7 +2960,8 @@ class LUAddNode(LogicalUnit):
         if (existing_node.primary_ip != primary_ip or
             existing_node.secondary_ip != secondary_ip):
           raise errors.OpPrereqError("Readded node doesn't have the same IP"
-                                     " address configuration as before")
+                                     " address configuration as before",
+                                     errors.ECODE_INVAL)
         continue
 
       if (existing_node.primary_ip == primary_ip or
@@ -2712,7 +2969,8 @@ class LUAddNode(LogicalUnit):
           existing_node.primary_ip == secondary_ip or
           existing_node.secondary_ip == secondary_ip):
         raise errors.OpPrereqError("New node ip address(es) conflict with"
-                                   " existing node %s" % existing_node.name)
+                                   " existing node %s" % existing_node.name,
+                                   errors.ECODE_NOTUNIQUE)
 
     # check that the type of the node (single versus dual homed) is the
     # same as for the master
@@ -2722,31 +2980,32 @@ class LUAddNode(LogicalUnit):
     if master_singlehomed != newbie_singlehomed:
       if master_singlehomed:
         raise errors.OpPrereqError("The master has no private ip but the"
-                                   " new node has one")
+                                   " new node has one",
+                                   errors.ECODE_INVAL)
       else:
         raise errors.OpPrereqError("The master has a private ip but the"
-                                   " new node doesn't have one")
+                                   " new node doesn't have one",
+                                   errors.ECODE_INVAL)
 
     # checks reachability
     if not utils.TcpPing(primary_ip, constants.DEFAULT_NODED_PORT):
-      raise errors.OpPrereqError("Node not reachable by ping")
+      raise errors.OpPrereqError("Node not reachable by ping",
+                                 errors.ECODE_ENVIRON)
 
     if not newbie_singlehomed:
       # check reachability from my secondary ip to newbie's secondary ip
       if not utils.TcpPing(secondary_ip, constants.DEFAULT_NODED_PORT,
                            source=myself.secondary_ip):
         raise errors.OpPrereqError("Node secondary ip not reachable by TCP"
-                                   " based ping to noded port")
+                                   " based ping to noded port",
+                                   errors.ECODE_ENVIRON)
 
-    cp_size = self.cfg.GetClusterInfo().candidate_pool_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.master_candidate = _DecideSelfPromotion(self, exceptions=exceptions)
 
     if self.op.readd:
       self.new_node = self.cfg.GetNodeInfo(node)
@@ -2770,7 +3029,7 @@ class LUAddNode(LogicalUnit):
     # 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
+      new_node.drained = new_node.offline = False # pylint: disable-msg=W0201
       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
@@ -2791,20 +3050,21 @@ class LUAddNode(LogicalUnit):
                                (constants.PROTOCOL_VERSION, result.payload))
 
     # setup ssh on node
-    logging.info("Copy ssh key to node %s", node)
-    priv_key, pub_key, _ = ssh.GetUserFiles(constants.GANETI_RUNAS)
-    keyarray = []
-    keyfiles = [constants.SSH_HOST_DSA_PRIV, constants.SSH_HOST_DSA_PUB,
-                constants.SSH_HOST_RSA_PRIV, constants.SSH_HOST_RSA_PUB,
-                priv_key, pub_key]
-
-    for i in keyfiles:
-      keyarray.append(utils.ReadFile(i))
-
-    result = self.rpc.call_node_add(node, keyarray[0], keyarray[1],
-                                    keyarray[2],
-                                    keyarray[3], keyarray[4], keyarray[5])
-    result.Raise("Cannot transfer ssh keys to the new node")
+    if self.cfg.GetClusterInfo().modify_ssh_setup:
+      logging.info("Copy ssh key to node %s", node)
+      priv_key, pub_key, _ = ssh.GetUserFiles(constants.GANETI_RUNAS)
+      keyarray = []
+      keyfiles = [constants.SSH_HOST_DSA_PRIV, constants.SSH_HOST_DSA_PUB,
+                  constants.SSH_HOST_RSA_PRIV, constants.SSH_HOST_RSA_PUB,
+                  priv_key, pub_key]
+
+      for i in keyfiles:
+        keyarray.append(utils.ReadFile(i))
+
+      result = self.rpc.call_node_add(node, keyarray[0], keyarray[1],
+                                      keyarray[2], keyarray[3], keyarray[4],
+                                      keyarray[5])
+      result.Raise("Cannot transfer ssh keys to the new node")
 
     # Add node to our /etc/hosts, and add key to known_hosts
     if self.cfg.GetClusterInfo().modify_etc_hosts:
@@ -2814,7 +3074,7 @@ class LUAddNode(LogicalUnit):
       result = self.rpc.call_node_has_ip_address(new_node.name,
                                                  new_node.secondary_ip)
       result.Raise("Failure checking secondary ip on node %s" % new_node.name,
-                   prereq=True)
+                   prereq=True, ecode=errors.ECODE_ENVIRON)
       if not result.payload:
         raise errors.OpExecError("Node claims it doesn't have the secondary ip"
                                  " you gave (%s). Please fix and re-run this"
@@ -2822,7 +3082,7 @@ class LUAddNode(LogicalUnit):
 
     node_verify_list = [self.cfg.GetMasterNode()]
     node_verify_param = {
-      'nodelist': [node],
+      constants.NV_NODELIST: [node],
       # TODO: do a node-net-test as well?
     }
 
@@ -2830,10 +3090,11 @@ class LUAddNode(LogicalUnit):
                                        self.cfg.GetClusterName())
     for verifier in node_verify_list:
       result[verifier].Raise("Cannot communicate with node %s" % verifier)
-      nl_payload = result[verifier].payload['nodelist']
+      nl_payload = result[verifier].payload[constants.NV_NODELIST]
       if nl_payload:
         for failed in nl_payload:
-          feedback_fn("ssh/hostname verification failed %s -> %s" %
+          feedback_fn("ssh/hostname verification failed"
+                      " (checking from %s): %s" %
                       (verifier, nl_payload[failed]))
         raise errors.OpExecError("ssh/hostname verification failed.")
 
@@ -2841,17 +3102,17 @@ class LUAddNode(LogicalUnit):
       _RedistributeAncillaryFiles(self)
       self.context.ReaddNode(new_node)
       # make sure we redistribute the config
-      self.cfg.Update(new_node)
+      self.cfg.Update(new_node, feedback_fn)
       # 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()
+        msg = result.fail_msg
         if msg:
           self.LogWarning("Node failed to demote itself from master"
                           " candidate status: %s" % msg)
     else:
       _RedistributeAncillaryFiles(self, additional_nodes=[node])
-      self.context.AddNode(new_node)
+      self.context.AddNode(new_node, self.proc.GetECId())
 
 
 class LUSetNodeParams(LogicalUnit):
@@ -2866,17 +3127,20 @@ class LUSetNodeParams(LogicalUnit):
   def CheckArguments(self):
     node_name = self.cfg.ExpandNodeName(self.op.node_name)
     if node_name is None:
-      raise errors.OpPrereqError("Invalid node name '%s'" % self.op.node_name)
+      raise errors.OpPrereqError("Invalid node name '%s'" % self.op.node_name,
+                                 errors.ECODE_INVAL)
     self.op.node_name = node_name
     _CheckBooleanOpField(self.op, 'master_candidate')
     _CheckBooleanOpField(self.op, 'offline')
     _CheckBooleanOpField(self.op, 'drained')
     all_mods = [self.op.offline, self.op.master_candidate, self.op.drained]
     if all_mods.count(None) == 3:
-      raise errors.OpPrereqError("Please pass at least one modification")
+      raise errors.OpPrereqError("Please pass at least one modification",
+                                 errors.ECODE_INVAL)
     if all_mods.count(True) > 1:
       raise errors.OpPrereqError("Can't set the node into more than one"
-                                 " state at the same time")
+                                 " state at the same time",
+                                 errors.ECODE_INVAL)
 
   def ExpandNames(self):
     self.needed_locks = {locking.LEVEL_NODE: self.op.node_name}
@@ -2911,25 +3175,42 @@ class LUSetNodeParams(LogicalUnit):
       # 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")
+                                   " only via masterfailover",
+                                   errors.ECODE_INVAL)
+
+    # Boolean value that tells us whether we're offlining or draining the node
+    offline_or_drain = self.op.offline == True or self.op.drained == True
+    deoffline_or_drain = self.op.offline == False or self.op.drained == False
 
-    if ((self.op.master_candidate == False or self.op.offline == True or
-         self.op.drained == True) and node.master_candidate):
+    if (node.master_candidate and
+        (self.op.master_candidate == False or offline_or_drain)):
       cp_size = self.cfg.GetClusterInfo().candidate_pool_size
-      num_candidates, _ = self.cfg.GetMasterCandidateStats()
-      if num_candidates <= cp_size:
+      mc_now, mc_should, mc_max = self.cfg.GetMasterCandidateStats()
+      if mc_now <= cp_size:
         msg = ("Not enough master candidates (desired"
-               " %d, new value will be %d)" % (cp_size, num_candidates-1))
-        if self.op.force:
+               " %d, new value will be %d)" % (cp_size, mc_now-1))
+        # Only allow forcing the operation if it's an offline/drain operation,
+        # and we could not possibly promote more nodes.
+        # FIXME: this can still lead to issues if in any way another node which
+        # could be promoted appears in the meantime.
+        if self.op.force and offline_or_drain and mc_should == mc_max:
           self.LogWarning(msg)
         else:
-          raise errors.OpPrereqError(msg)
+          raise errors.OpPrereqError(msg, errors.ECODE_INVAL)
 
     if (self.op.master_candidate == True and
         ((node.offline and not self.op.offline == False) or
          (node.drained and not self.op.drained == False))):
       raise errors.OpPrereqError("Node '%s' is offline or drained, can't set"
-                                 " to master_candidate" % node.name)
+                                 " to master_candidate" % node.name,
+                                 errors.ECODE_INVAL)
+
+    # If we're being deofflined/drained, we'll MC ourself if needed
+    if (deoffline_or_drain and not offline_or_drain and not
+        self.op.master_candidate == True and not node.master_candidate):
+      self.op.master_candidate = _DecideSelfPromotion(self)
+      if self.op.master_candidate:
+        self.LogInfo("Autopromoting node to master candidate")
 
     return
 
@@ -2973,7 +3254,7 @@ class LUSetNodeParams(LogicalUnit):
           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()
+          msg = rrc.fail_msg
           if msg:
             self.LogWarning("Node failed to demote itself: %s" % msg)
         if node.offline:
@@ -2981,7 +3262,7 @@ class LUSetNodeParams(LogicalUnit):
           result.append(("offline", "clear offline status due to drain"))
 
     # this will trigger configuration file update, if needed
-    self.cfg.Update(node)
+    self.cfg.Update(node, feedback_fn)
     # this will trigger job queue propagation or cleanup
     if changed_mc:
       self.context.ReaddNode(node)
@@ -2999,11 +3280,13 @@ class LUPowercycleNode(NoHooksLU):
   def CheckArguments(self):
     node_name = self.cfg.ExpandNodeName(self.op.node_name)
     if node_name is None:
-      raise errors.OpPrereqError("Invalid node name '%s'" % self.op.node_name)
+      raise errors.OpPrereqError("Invalid node name '%s'" % self.op.node_name,
+                                 errors.ECODE_NOENT)
     self.op.node_name = node_name
     if node_name == self.cfg.GetMasterNode() and not self.op.force:
       raise errors.OpPrereqError("The node is the master and the force"
-                                 " parameter was not set")
+                                 " parameter was not set",
+                                 errors.ECODE_INVAL)
 
   def ExpandNames(self):
     """Locking for PowercycleNode.
@@ -3074,6 +3357,7 @@ class LUQueryClusterInfo(NoHooksLU):
       "file_storage_dir": cluster.file_storage_dir,
       "ctime": cluster.ctime,
       "mtime": cluster.mtime,
+      "uuid": cluster.uuid,
       "tags": list(cluster.GetTags()),
       }
 
@@ -3219,6 +3503,8 @@ def _AssembleInstanceDisks(lu, instance, ignore_secondaries=False,
 
   # 2nd pass, do only the primary node
   for inst_disk in instance.disks:
+    dev_path = None
+
     for node, node_disk in inst_disk.ComputeNodeTree(instance.primary_node):
       if node != instance.primary_node:
         continue
@@ -3233,8 +3519,10 @@ def _AssembleInstanceDisks(lu, instance, ignore_secondaries=False,
                            " (is_primary=True, pass=2): %s",
                            inst_disk.iv_name, node, msg)
         disks_ok = False
-    device_info.append((instance.primary_node, inst_disk.iv_name,
-                        result.payload))
+      else:
+        dev_path = result.payload
+
+    device_info.append((instance.primary_node, inst_disk.iv_name, dev_path))
 
   # leave the disks configured for the primary node
   # this is a workaround that would be fixed better by
@@ -3358,15 +3646,18 @@ def _CheckNodeFreeMemory(lu, node, reason, requested, hypervisor_name):
 
   """
   nodeinfo = lu.rpc.call_node_info([node], lu.cfg.GetVGName(), hypervisor_name)
-  nodeinfo[node].Raise("Can't get data from node %s" % node, prereq=True)
+  nodeinfo[node].Raise("Can't get data from node %s" % node,
+                       prereq=True, ecode=errors.ECODE_ENVIRON)
   free_mem = nodeinfo[node].payload.get('memory_free', None)
   if not isinstance(free_mem, int):
     raise errors.OpPrereqError("Can't compute free memory on node %s, result"
-                               " was '%s'" % (node, free_mem))
+                               " was '%s'" % (node, free_mem),
+                               errors.ECODE_ENVIRON)
   if requested > free_mem:
     raise errors.OpPrereqError("Not enough memory on node %s for %s:"
                                " needed %s MiB, available %s MiB" %
-                               (node, reason, requested, free_mem))
+                               (node, reason, requested, free_mem),
+                               errors.ECODE_NORES)
 
 
 class LUStartupInstance(LogicalUnit):
@@ -3409,7 +3700,8 @@ class LUStartupInstance(LogicalUnit):
     if self.beparams:
       if not isinstance(self.beparams, dict):
         raise errors.OpPrereqError("Invalid beparams passed: %s, expected"
-                                   " dict" % (type(self.beparams), ))
+                                   " dict" % (type(self.beparams), ),
+                                   errors.ECODE_INVAL)
       # fill the beparams dict
       utils.ForceDictType(self.beparams, constants.BES_PARAMETER_TYPES)
       self.op.beparams = self.beparams
@@ -3419,7 +3711,8 @@ class LUStartupInstance(LogicalUnit):
     if self.hvparams:
       if not isinstance(self.hvparams, dict):
         raise errors.OpPrereqError("Invalid hvparams passed: %s, expected"
-                                   " dict" % (type(self.hvparams), ))
+                                   " dict" % (type(self.hvparams), ),
+                                   errors.ECODE_INVAL)
 
       # check hypervisor parameter syntax (locally)
       cluster = self.cfg.GetClusterInfo()
@@ -3442,7 +3735,7 @@ class LUStartupInstance(LogicalUnit):
                                               instance.name,
                                               instance.hypervisor)
     remote_info.Raise("Error checking node %s" % instance.primary_node,
-                      prereq=True)
+                      prereq=True, ecode=errors.ECODE_ENVIRON)
     if not remote_info.payload: # not running already
       _CheckNodeFreeMemory(self, instance.primary_node,
                            "starting instance %s" % instance.name,
@@ -3478,6 +3771,13 @@ 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,
@@ -3497,6 +3797,7 @@ 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)
@@ -3532,10 +3833,12 @@ 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)
@@ -3558,6 +3861,13 @@ class LUShutdownInstance(LogicalUnit):
   _OP_REQP = ["instance_name"]
   REQ_BGL = False
 
+  def CheckArguments(self):
+    """Check the arguments.
+
+    """
+    self.timeout = getattr(self.op, "timeout",
+                           constants.DEFAULT_SHUTDOWN_TIMEOUT)
+
   def ExpandNames(self):
     self._ExpandAndLockInstance()
 
@@ -3568,6 +3878,7 @@ class LUShutdownInstance(LogicalUnit):
 
     """
     env = _BuildInstanceHookEnvByObject(self, self.instance)
+    env["TIMEOUT"] = self.timeout
     nl = [self.cfg.GetMasterNode()] + list(self.instance.all_nodes)
     return env, nl, nl
 
@@ -3588,8 +3899,9 @@ class LUShutdownInstance(LogicalUnit):
     """
     instance = self.instance
     node_current = instance.primary_node
+    timeout = self.timeout
     self.cfg.MarkInstanceDown(instance.name)
-    result = self.rpc.call_instance_shutdown(node_current, instance)
+    result = self.rpc.call_instance_shutdown(node_current, instance, timeout)
     msg = result.fail_msg
     if msg:
       self.proc.LogWarning("Could not shutdown instance: %s" % msg)
@@ -3632,31 +3944,38 @@ class LUReinstallInstance(LogicalUnit):
 
     if instance.disk_template == constants.DT_DISKLESS:
       raise errors.OpPrereqError("Instance '%s' has no disks" %
-                                 self.op.instance_name)
+                                 self.op.instance_name,
+                                 errors.ECODE_INVAL)
     if instance.admin_up:
       raise errors.OpPrereqError("Instance '%s' is marked to be up" %
-                                 self.op.instance_name)
+                                 self.op.instance_name,
+                                 errors.ECODE_STATE)
     remote_info = self.rpc.call_instance_info(instance.primary_node,
                                               instance.name,
                                               instance.hypervisor)
     remote_info.Raise("Error checking node %s" % instance.primary_node,
-                      prereq=True)
+                      prereq=True, ecode=errors.ECODE_ENVIRON)
     if remote_info.payload:
       raise errors.OpPrereqError("Instance '%s' is running on the node %s" %
                                  (self.op.instance_name,
-                                  instance.primary_node))
+                                  instance.primary_node),
+                                 errors.ECODE_STATE)
 
     self.op.os_type = getattr(self.op, "os_type", None)
+    self.op.force_variant = getattr(self.op, "force_variant", False)
     if self.op.os_type is not None:
       # OS verification
       pnode = self.cfg.GetNodeInfo(
         self.cfg.ExpandNodeName(instance.primary_node))
       if pnode is None:
         raise errors.OpPrereqError("Primary node '%s' is unknown" %
-                                   self.op.pnode)
+                                   self.op.pnode, errors.ECODE_NOENT)
       result = self.rpc.call_os_get(pnode.name, self.op.os_type)
       result.Raise("OS '%s' not in supported OS list for primary node %s" %
-                   (self.op.os_type, pnode.name), prereq=True)
+                   (self.op.os_type, pnode.name),
+                   prereq=True, ecode=errors.ECODE_INVAL)
+      if not self.op.force_variant:
+        _CheckOSVariant(result.payload, self.op.os_type)
 
     self.instance = instance
 
@@ -3669,12 +3988,13 @@ class LUReinstallInstance(LogicalUnit):
     if self.op.os_type is not None:
       feedback_fn("Changing OS to '%s'..." % self.op.os_type)
       inst.os = self.op.os_type
-      self.cfg.Update(inst)
+      self.cfg.Update(inst, feedback_fn)
 
     _StartInstanceDisks(self, inst, None)
     try:
       feedback_fn("Running the instance OS create scripts...")
-      result = self.rpc.call_instance_os_add(inst.primary_node, inst, True)
+      # FIXME: pass debug option from opcode to backend
+      result = self.rpc.call_instance_os_add(inst.primary_node, inst, True, 0)
       result.Raise("Could not install OS for instance %s on node %s" %
                    (inst.name, inst.primary_node))
     finally:
@@ -3695,12 +4015,12 @@ class LURecreateInstanceDisks(LogicalUnit):
 
     """
     if not isinstance(self.op.disks, list):
-      raise errors.OpPrereqError("Invalid disks parameter")
+      raise errors.OpPrereqError("Invalid disks parameter", errors.ECODE_INVAL)
     for item in self.op.disks:
       if (not isinstance(item, int) or
           item < 0):
         raise errors.OpPrereqError("Invalid disk specification '%s'" %
-                                   str(item))
+                                   str(item), errors.ECODE_INVAL)
 
   def ExpandNames(self):
     self._ExpandAndLockInstance()
@@ -3728,26 +4048,27 @@ class LURecreateInstanceDisks(LogicalUnit):
 
     if instance.disk_template == constants.DT_DISKLESS:
       raise errors.OpPrereqError("Instance '%s' has no disks" %
-                                 self.op.instance_name)
+                                 self.op.instance_name, errors.ECODE_INVAL)
     if instance.admin_up:
       raise errors.OpPrereqError("Instance '%s' is marked to be up" %
-                                 self.op.instance_name)
+                                 self.op.instance_name, errors.ECODE_STATE)
     remote_info = self.rpc.call_instance_info(instance.primary_node,
                                               instance.name,
                                               instance.hypervisor)
     remote_info.Raise("Error checking node %s" % instance.primary_node,
-                      prereq=True)
+                      prereq=True, ecode=errors.ECODE_ENVIRON)
     if remote_info.payload:
       raise errors.OpPrereqError("Instance '%s' is running on the node %s" %
                                  (self.op.instance_name,
-                                  instance.primary_node))
+                                  instance.primary_node), errors.ECODE_STATE)
 
     if not self.op.disks:
       self.op.disks = range(len(instance.disks))
     else:
       for idx in self.op.disks:
         if idx >= len(instance.disks):
-          raise errors.OpPrereqError("Invalid disk index passed '%s'" % idx)
+          raise errors.OpPrereqError("Invalid disk index passed '%s'" % idx,
+                                     errors.ECODE_INVAL)
 
     self.instance = instance
 
@@ -3756,7 +4077,7 @@ class LURecreateInstanceDisks(LogicalUnit):
 
     """
     to_skip = []
-    for idx, disk in enumerate(self.instance.disks):
+    for idx, _ in enumerate(self.instance.disks):
       if idx not in self.op.disks: # disk idx has not been passed in
         to_skip.append(idx)
         continue
@@ -3793,36 +4114,37 @@ class LURenameInstance(LogicalUnit):
       self.cfg.ExpandInstanceName(self.op.instance_name))
     if instance is None:
       raise errors.OpPrereqError("Instance '%s' not known" %
-                                 self.op.instance_name)
+                                 self.op.instance_name, errors.ECODE_NOENT)
     _CheckNodeOnline(self, instance.primary_node)
 
     if instance.admin_up:
       raise errors.OpPrereqError("Instance '%s' is marked to be up" %
-                                 self.op.instance_name)
+                                 self.op.instance_name, errors.ECODE_STATE)
     remote_info = self.rpc.call_instance_info(instance.primary_node,
                                               instance.name,
                                               instance.hypervisor)
     remote_info.Raise("Error checking node %s" % instance.primary_node,
-                      prereq=True)
+                      prereq=True, ecode=errors.ECODE_ENVIRON)
     if remote_info.payload:
       raise errors.OpPrereqError("Instance '%s' is running on the node %s" %
                                  (self.op.instance_name,
-                                  instance.primary_node))
+                                  instance.primary_node), errors.ECODE_STATE)
     self.instance = instance
 
     # new name verification
-    name_info = utils.HostInfo(self.op.new_name)
+    name_info = utils.GetHostInfo(self.op.new_name)
 
     self.op.new_name = new_name = name_info.name
     instance_list = self.cfg.GetInstanceList()
     if new_name in instance_list:
       raise errors.OpPrereqError("Instance '%s' is already in the cluster" %
-                                 new_name)
+                                 new_name, errors.ECODE_EXISTS)
 
     if not getattr(self.op, "ignore_ip", False):
       if utils.TcpPing(name_info.ip, constants.DEFAULT_NODED_PORT):
         raise errors.OpPrereqError("IP %s of instance %s already in use" %
-                                   (name_info.ip, new_name))
+                                   (name_info.ip, new_name),
+                                   errors.ECODE_NOTUNIQUE)
 
 
   def Exec(self, feedback_fn):
@@ -3856,7 +4178,7 @@ class LURenameInstance(LogicalUnit):
     _StartInstanceDisks(self, inst, None)
     try:
       result = self.rpc.call_instance_run_rename(inst.primary_node, inst,
-                                                 old_name)
+                                                 old_name, 0)
       msg = result.fail_msg
       if msg:
         msg = ("Could not run OS rename script for instance %s on node %s"
@@ -3876,6 +4198,13 @@ 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] = []
@@ -3892,6 +4221,7 @@ class LURemoveInstance(LogicalUnit):
 
     """
     env = _BuildInstanceHookEnvByObject(self, self.instance)
+    env["SHUTDOWN_TIMEOUT"] = self.shutdown_timeout
     nl = [self.cfg.GetMasterNode()]
     return env, nl, nl
 
@@ -3913,7 +4243,8 @@ 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:
@@ -3941,8 +4272,11 @@ class LUQueryInstances(NoHooksLU):
   """Logical unit for querying instances.
 
   """
+  # pylint: disable-msg=W0142
   _OP_REQP = ["output_fields", "names", "use_locking"]
   REQ_BGL = False
+  _SIMPLE_FIELDS = ["name", "os", "network_port", "hypervisor",
+                    "serial_no", "ctime", "mtime", "uuid"]
   _FIELDS_STATIC = utils.FieldSet(*["name", "os", "pnode", "snodes",
                                     "admin_state",
                                     "disk_template", "ip", "mac", "bridge",
@@ -3955,11 +4289,11 @@ class LUQueryInstances(NoHooksLU):
                                     r"(nic)\.(bridge)/([0-9]+)",
                                     r"(nic)\.(macs|ips|modes|links|bridges)",
                                     r"(disk|nic)\.(count)",
-                                    "serial_no", "hypervisor", "hvparams",
-                                    "ctime", "mtime",
-                                    ] +
+                                    "hvparams",
+                                    ] + _SIMPLE_FIELDS +
                                   ["hv/%s" % name
-                                   for name in constants.HVS_PARAMETERS] +
+                                   for name in constants.HVS_PARAMETERS
+                                   if name not in constants.HVC_GLOBALS] +
                                   ["be/%s" % name
                                    for name in constants.BES_PARAMETERS])
   _FIELDS_DYNAMIC = utils.FieldSet("oper_state", "oper_ram", "status")
@@ -4000,6 +4334,8 @@ class LUQueryInstances(NoHooksLU):
     """Computes the list of nodes and their attributes.
 
     """
+    # pylint: disable-msg=R0912
+    # way too many branches here
     all_info = self.cfg.GetAllInstancesInfo()
     if self.wanted == locking.ALL_SET:
       # caller didn't specify instance names, so ordering is not important
@@ -4037,7 +4373,7 @@ class LUQueryInstances(NoHooksLU):
         if result.offline:
           # offline nodes will be in both lists
           off_nodes.append(name)
-        if result.failed or result.fail_msg:
+        if result.fail_msg:
           bad_nodes.append(name)
         else:
           if result.payload:
@@ -4054,16 +4390,14 @@ class LUQueryInstances(NoHooksLU):
     cluster = self.cfg.GetClusterInfo()
     for instance in instance_list:
       iout = []
-      i_hv = cluster.FillHV(instance)
+      i_hv = cluster.FillHV(instance, skip_globals=True)
       i_be = cluster.FillBE(instance)
       i_nicp = [objects.FillDict(cluster.nicparams[constants.PP_DEFAULT],
                                  nic.nicparams) for nic in instance.nics]
       for field in self.op.output_fields:
         st_match = self._FIELDS_STATIC.Matches(field)
-        if field == "name":
-          val = instance.name
-        elif field == "os":
-          val = instance.os
+        if field in self._SIMPLE_FIELDS:
+          val = getattr(instance, field)
         elif field == "pnode":
           val = instance.primary_node
         elif field == "snodes":
@@ -4140,20 +4474,11 @@ class LUQueryInstances(NoHooksLU):
           val = _ComputeDiskSize(instance.disk_template, disk_sizes)
         elif field == "tags":
           val = list(instance.GetTags())
-        elif field == "serial_no":
-          val = instance.serial_no
-        elif field == "ctime":
-          val = instance.ctime
-        elif field == "mtime":
-          val = instance.mtime
-        elif field == "network_port":
-          val = instance.network_port
-        elif field == "hypervisor":
-          val = instance.hypervisor
         elif field == "hvparams":
           val = i_hv
         elif (field.startswith(HVPREFIX) and
-              field[len(HVPREFIX):] in constants.HVS_PARAMETERS):
+              field[len(HVPREFIX):] in constants.HVS_PARAMETERS and
+              field[len(HVPREFIX):] not in constants.HVC_GLOBALS):
           val = i_hv.get(field[len(HVPREFIX):], None)
         elif field == "beparams":
           val = i_be
@@ -4235,6 +4560,13 @@ 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] = []
@@ -4252,6 +4584,7 @@ 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)
@@ -4270,7 +4603,8 @@ class LUFailoverInstance(LogicalUnit):
     bep = self.cfg.GetClusterInfo().FillBE(instance)
     if instance.disk_template not in constants.DTS_NET_MIRROR:
       raise errors.OpPrereqError("Instance's disk layout is not"
-                                 " network mirrored, cannot failover.")
+                                 " network mirrored, cannot failover.",
+                                 errors.ECODE_STATE)
 
     secondary_nodes = instance.secondary_nodes
     if not secondary_nodes:
@@ -4304,19 +4638,23 @@ class LUFailoverInstance(LogicalUnit):
     source_node = instance.primary_node
     target_node = instance.secondary_nodes[0]
 
-    feedback_fn("* checking disk consistency between source and target")
-    for dev in instance.disks:
-      # for drbd, these are drbd over lvm
-      if not _CheckDiskConsistency(self, dev, target_node, False):
-        if instance.admin_up and not self.op.ignore_consistency:
-          raise errors.OpExecError("Disk %s is degraded on target node,"
-                                   " aborting failover." % dev.iv_name)
+    if instance.admin_up:
+      feedback_fn("* checking disk consistency between source and target")
+      for dev in instance.disks:
+        # for drbd, these are drbd over lvm
+        if not _CheckDiskConsistency(self, dev, target_node, False):
+          if not self.op.ignore_consistency:
+            raise errors.OpExecError("Disk %s is degraded on target node,"
+                                     " aborting failover." % dev.iv_name)
+    else:
+      feedback_fn("* not checking disk consistency as instance is not running")
 
     feedback_fn("* shutting down instance on source node")
     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:
@@ -4335,7 +4673,7 @@ class LUFailoverInstance(LogicalUnit):
 
     instance.primary_node = target_node
     # distribute new instance config to the other nodes
-    self.cfg.Update(instance)
+    self.cfg.Update(instance, feedback_fn)
 
     # Only start the instance if it's marked as up
     if instance.admin_up:
@@ -4408,12 +4746,19 @@ 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)
     if target_node is None:
       raise errors.OpPrereqError("Node '%s' not known" %
-                                  self.op.target_node)
+                                  self.op.target_node, errors.ECODE_NOENT)
     self.op.target_node = target_node
     self.needed_locks[locking.LEVEL_NODE] = [target_node]
     self.recalculate_locks[locking.LEVEL_NODE] = constants.LOCKS_APPEND
@@ -4430,6 +4775,7 @@ 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,
@@ -4454,14 +4800,15 @@ class LUMoveInstance(LogicalUnit):
 
     if target_node == instance.primary_node:
       raise errors.OpPrereqError("Instance %s is already on the node %s" %
-                                 (instance.name, target_node))
+                                 (instance.name, target_node),
+                                 errors.ECODE_STATE)
 
     bep = self.cfg.GetClusterInfo().FillBE(instance)
 
     for idx, dsk in enumerate(instance.disks):
       if dsk.dev_type not in (constants.LD_LV, constants.LD_FILE):
         raise errors.OpPrereqError("Instance disk %d has a complex layout,"
-                                   " cannot copy")
+                                   " cannot copy" % idx, errors.ECODE_STATE)
 
     _CheckNodeOnline(self, target_node)
     _CheckNodeNotDrained(self, target_node)
@@ -4493,7 +4840,8 @@ 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:
@@ -4550,7 +4898,7 @@ class LUMoveInstance(LogicalUnit):
                                  (",".join(errs),))
 
     instance.primary_node = target_node
-    self.cfg.Update(instance)
+    self.cfg.Update(instance, feedback_fn)
 
     self.LogInfo("Removing the disks on the original node")
     _RemoveDisks(self, instance, target_node=source_node)
@@ -4586,7 +4934,8 @@ class LUMigrateNode(LogicalUnit):
   def ExpandNames(self):
     self.op.node_name = self.cfg.ExpandNodeName(self.op.node_name)
     if self.op.node_name is None:
-      raise errors.OpPrereqError("Node '%s' not known" % self.op.node_name)
+      raise errors.OpPrereqError("Node '%s' not known" % self.op.node_name,
+                                 errors.ECODE_NOENT)
 
     self.needed_locks = {
       locking.LEVEL_NODE: [self.op.node_name],
@@ -4650,11 +4999,11 @@ class TLMigrateInstance(Tasklet):
       self.cfg.ExpandInstanceName(self.instance_name))
     if instance is None:
       raise errors.OpPrereqError("Instance '%s' not known" %
-                                 self.instance_name)
+                                 self.instance_name, errors.ECODE_NOENT)
 
     if instance.disk_template != constants.DT_DRBD8:
       raise errors.OpPrereqError("Instance's disk layout is not"
-                                 " drbd8, cannot migrate.")
+                                 " drbd8, cannot migrate.", errors.ECODE_STATE)
 
     secondary_nodes = instance.secondary_nodes
     if not secondary_nodes:
@@ -4676,7 +5025,8 @@ class TLMigrateInstance(Tasklet):
       _CheckNodeNotDrained(self, target_node)
       result = self.rpc.call_instance_migratable(instance.primary_node,
                                                  instance)
-      result.Raise("Can't migrate, please use failover", prereq=True)
+      result.Raise("Can't migrate, please use failover",
+                   prereq=True, ecode=errors.ECODE_STATE)
 
     self.instance = instance
 
@@ -4788,7 +5138,7 @@ class TLMigrateInstance(Tasklet):
       self.feedback_fn("* instance running on secondary node (%s),"
                        " updating config" % target_node)
       instance.primary_node = target_node
-      self.cfg.Update(instance)
+      self.cfg.Update(instance, self.feedback_fn)
       demoted_node = source_node
     else:
       self.feedback_fn("* instance confirmed to be running on its"
@@ -4838,8 +5188,8 @@ class TLMigrateInstance(Tasklet):
                                                     False)
     abort_msg = abort_result.fail_msg
     if abort_msg:
-      logging.error("Aborting migration failed on target node %s: %s" %
-                    (target_node, abort_msg))
+      logging.error("Aborting migration failed on target node %s: %s",
+                    target_node, abort_msg)
       # Don't raise an exception here, as we stil have to try to revert the
       # disk status, even if this step failed.
 
@@ -4893,6 +5243,7 @@ class TLMigrateInstance(Tasklet):
     if msg:
       logging.error("Instance pre-migration failed, trying to revert"
                     " disk status: %s", msg)
+      self.feedback_fn("Pre-migration failed, aborting")
       self._AbortMigration()
       self._RevertDiskStatus()
       raise errors.OpExecError("Could not pre-migrate instance %s: %s" %
@@ -4907,6 +5258,7 @@ class TLMigrateInstance(Tasklet):
     if msg:
       logging.error("Instance migration failed, trying to revert"
                     " disk status: %s", msg)
+      self.feedback_fn("Migration failed, aborting")
       self._AbortMigration()
       self._RevertDiskStatus()
       raise errors.OpExecError("Could not migrate instance %s: %s" %
@@ -4915,7 +5267,7 @@ class TLMigrateInstance(Tasklet):
 
     instance.primary_node = target_node
     # distribute new instance config to the other nodes
-    self.cfg.Update(instance)
+    self.cfg.Update(instance, self.feedback_fn)
 
     result = self.rpc.call_finalize_migration(target_node,
                                               instance,
@@ -4924,7 +5276,7 @@ class TLMigrateInstance(Tasklet):
     msg = result.fail_msg
     if msg:
       logging.error("Instance migration succeeded, but finalization failed:"
-                    " %s" % msg)
+                    " %s", msg)
       raise errors.OpExecError("Could not finalize instance migration: %s" %
                                msg)
 
@@ -5038,7 +5390,7 @@ def _GenerateUniqueNames(lu, exts):
   """
   results = []
   for val in exts:
-    new_id = lu.cfg.GenerateUniqueID()
+    new_id = lu.cfg.GenerateUniqueID(lu.proc.GetECId())
     results.append("%s%s" % (new_id, val))
   return results
 
@@ -5050,7 +5402,7 @@ def _GenerateDRBD8Branch(lu, primary, secondary, size, names, iv_name,
   """
   port = lu.cfg.AllocatePort()
   vgname = lu.cfg.GetVGName()
-  shared_secret = lu.cfg.GenerateDRBDSecret()
+  shared_secret = lu.cfg.GenerateDRBDSecret(lu.proc.GetECId())
   dev_data = objects.Disk(dev_type=constants.LD_LV, size=size,
                           logical_id=(vgname, names[0]))
   dev_meta = objects.Disk(dev_type=constants.LD_LV, size=128,
@@ -5167,7 +5519,7 @@ def _CreateDisks(lu, instance, to_skip=None, target_node=None):
     result = lu.rpc.call_file_storage_dir_create(pnode, file_storage_dir)
 
     result.Raise("Failed to create directory '%s' on"
-                 " node %s: %s" % (file_storage_dir, pnode))
+                 " node %s" % (file_storage_dir, pnode))
 
   # Note: this needs to be kept in sync with adding of disks in
   # LUSetInstanceParams
@@ -5218,10 +5570,10 @@ def _RemoveDisks(lu, instance, target_node=None):
 
   if instance.disk_template == constants.DT_FILE:
     file_storage_dir = os.path.dirname(instance.disks[0].logical_id[1])
-    if target_node is node:
-      tgt = instance.primary_node
+    if target_node:
+      tgt = target_node
     else:
-      tgt = instance.target_node
+      tgt = instance.primary_node
     result = lu.rpc.call_file_storage_dir_remove(tgt, file_storage_dir)
     if result.fail_msg:
       lu.LogWarning("Could not remove directory '%s' on node %s: %s",
@@ -5290,13 +5642,26 @@ class LUCreateInstance(LogicalUnit):
               "hvparams", "beparams"]
   REQ_BGL = False
 
+  def CheckArguments(self):
+    """Check arguments.
+
+    """
+    # do not require name_check to ease forward/backward compatibility
+    # for tools
+    if not hasattr(self.op, "name_check"):
+      self.op.name_check = True
+    if self.op.ip_check and not self.op.name_check:
+      # TODO: make the ip check more flexible and not depend on the name check
+      raise errors.OpPrereqError("Cannot do ip checks without a name check",
+                                 errors.ECODE_INVAL)
+
   def _ExpandNode(self, node):
     """Expands and checks one node name.
 
     """
     node_full = self.cfg.ExpandNodeName(node)
     if node_full is None:
-      raise errors.OpPrereqError("Unknown node %s" % node)
+      raise errors.OpPrereqError("Unknown node %s" % node, errors.ECODE_NOENT)
     return node_full
 
   def ExpandNames(self):
@@ -5318,11 +5683,12 @@ class LUCreateInstance(LogicalUnit):
     if self.op.mode not in (constants.INSTANCE_CREATE,
                             constants.INSTANCE_IMPORT):
       raise errors.OpPrereqError("Invalid instance creation mode '%s'" %
-                                 self.op.mode)
+                                 self.op.mode, errors.ECODE_INVAL)
 
     # disk template and mirror node verification
     if self.op.disk_template not in constants.DISK_TEMPLATES:
-      raise errors.OpPrereqError("Invalid disk template name")
+      raise errors.OpPrereqError("Invalid disk template name",
+                                 errors.ECODE_INVAL)
 
     if self.op.hypervisor is None:
       self.op.hypervisor = self.cfg.GetHypervisorType()
@@ -5332,7 +5698,8 @@ class LUCreateInstance(LogicalUnit):
     if self.op.hypervisor not in enabled_hvs:
       raise errors.OpPrereqError("Selected hypervisor (%s) not enabled in the"
                                  " cluster (%s)" % (self.op.hypervisor,
-                                  ",".join(enabled_hvs)))
+                                  ",".join(enabled_hvs)),
+                                 errors.ECODE_STATE)
 
     # check hypervisor parameter syntax (locally)
     utils.ForceDictType(self.op.hvparams, constants.HVS_PARAMETER_TYPES)
@@ -5341,6 +5708,8 @@ class LUCreateInstance(LogicalUnit):
     hv_type = hypervisor.GetHypervisor(self.op.hypervisor)
     hv_type.CheckParameterSyntax(filled_hvp)
     self.hv_full = filled_hvp
+    # check that we don't specify global parameters on an instance
+    _CheckGlobalHvParams(self.op.hvparams)
 
     # fill and remember the beparams dict
     utils.ForceDictType(self.op.beparams, constants.BES_PARAMETER_TYPES)
@@ -5350,14 +5719,20 @@ class LUCreateInstance(LogicalUnit):
     #### instance parameters check
 
     # instance name verification
-    hostname1 = utils.HostInfo(self.op.instance_name)
-    self.op.instance_name = instance_name = hostname1.name
+    if self.op.name_check:
+      hostname1 = utils.GetHostInfo(self.op.instance_name)
+      self.op.instance_name = instance_name = hostname1.name
+      # used in CheckPrereq for ip ping check
+      self.check_ip = hostname1.ip
+    else:
+      instance_name = self.op.instance_name
+      self.check_ip = None
 
     # this is just a preventive check, but someone might still add this
     # instance in the meantime, and creation will fail at lock-add time
     if instance_name in self.cfg.GetInstanceList():
       raise errors.OpPrereqError("Instance '%s' is already in the cluster" %
-                                 instance_name)
+                                 instance_name, errors.ECODE_EXISTS)
 
     self.add_locks[locking.LEVEL_INSTANCE] = instance_name
 
@@ -5380,37 +5755,44 @@ class LUCreateInstance(LogicalUnit):
       if ip is None or ip.lower() == constants.VALUE_NONE:
         nic_ip = None
       elif ip.lower() == constants.VALUE_AUTO:
+        if not self.op.name_check:
+          raise errors.OpPrereqError("IP address set to auto but name checks"
+                                     " have been skipped. Aborting.",
+                                     errors.ECODE_INVAL)
         nic_ip = hostname1.ip
       else:
         if not utils.IsValidIP(ip):
           raise errors.OpPrereqError("Given IP address '%s' doesn't look"
-                                     " like a valid IP" % ip)
+                                     " like a valid IP" % ip,
+                                     errors.ECODE_INVAL)
         nic_ip = ip
 
-      # TODO: check the ip for uniqueness !!
+      # TODO: check the ip address for uniqueness
       if nic_mode == constants.NIC_MODE_ROUTED and not nic_ip:
-        raise errors.OpPrereqError("Routed nic mode requires an ip address")
+        raise errors.OpPrereqError("Routed nic mode requires an ip address",
+                                   errors.ECODE_INVAL)
 
       # MAC address verification
       mac = nic.get("mac", constants.VALUE_AUTO)
       if mac not in (constants.VALUE_AUTO, constants.VALUE_GENERATE):
-        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)
+        mac = utils.NormalizeAndValidateMac(mac)
+
+        try:
+          self.cfg.ReserveMAC(mac, self.proc.GetECId())
+        except errors.ReservationError:
+          raise errors.OpPrereqError("MAC address %s already in use"
+                                     " in cluster" % mac,
+                                     errors.ECODE_NOTUNIQUE)
 
       # bridge verification
       bridge = nic.get("bridge", None)
       link = nic.get("link", None)
       if bridge and link:
         raise errors.OpPrereqError("Cannot pass 'bridge' and 'link'"
-                                   " at the same time")
+                                   " at the same time", errors.ECODE_INVAL)
       elif bridge and nic_mode == constants.NIC_MODE_ROUTED:
-        raise errors.OpPrereqError("Cannot pass 'bridge' on a routed nic")
+        raise errors.OpPrereqError("Cannot pass 'bridge' on a routed nic",
+                                   errors.ECODE_INVAL)
       elif bridge:
         link = bridge
 
@@ -5431,32 +5813,32 @@ class LUCreateInstance(LogicalUnit):
       mode = disk.get("mode", constants.DISK_RDWR)
       if mode not in constants.DISK_ACCESS_SET:
         raise errors.OpPrereqError("Invalid disk access mode '%s'" %
-                                   mode)
+                                   mode, errors.ECODE_INVAL)
       size = disk.get("size", None)
       if size is None:
-        raise errors.OpPrereqError("Missing disk size")
+        raise errors.OpPrereqError("Missing disk size", errors.ECODE_INVAL)
       try:
         size = int(size)
-      except ValueError:
-        raise errors.OpPrereqError("Invalid disk size '%s'" % size)
+      except (TypeError, ValueError):
+        raise errors.OpPrereqError("Invalid disk size '%s'" % size,
+                                   errors.ECODE_INVAL)
       self.disks.append({"size": size, "mode": mode})
 
-    # used in CheckPrereq for ip ping check
-    self.check_ip = hostname1.ip
-
     # file storage checks
     if (self.op.file_driver and
         not self.op.file_driver in constants.FILE_DRIVER):
       raise errors.OpPrereqError("Invalid file driver name '%s'" %
-                                 self.op.file_driver)
+                                 self.op.file_driver, errors.ECODE_INVAL)
 
     if self.op.file_storage_dir and os.path.isabs(self.op.file_storage_dir):
-      raise errors.OpPrereqError("File storage directory path not absolute")
+      raise errors.OpPrereqError("File storage directory path not absolute",
+                                 errors.ECODE_INVAL)
 
     ### Node/iallocator related checks
     if [self.op.iallocator, self.op.pnode].count(None) != 1:
       raise errors.OpPrereqError("One and only one of iallocator and primary"
-                                 " node must be given")
+                                 " node must be given",
+                                 errors.ECODE_INVAL)
 
     if self.op.iallocator:
       self.needed_locks[locking.LEVEL_NODE] = locking.ALL_SET
@@ -5481,7 +5863,8 @@ class LUCreateInstance(LogicalUnit):
         self.op.src_node = None
         if os.path.isabs(src_path):
           raise errors.OpPrereqError("Importing an instance from an absolute"
-                                     " path requires a source node option.")
+                                     " path requires a source node option.",
+                                     errors.ECODE_INVAL)
       else:
         self.op.src_node = src_node = self._ExpandNode(src_node)
         if self.needed_locks[locking.LEVEL_NODE] is not locking.ALL_SET:
@@ -5490,9 +5873,16 @@ class LUCreateInstance(LogicalUnit):
           self.op.src_path = src_path = \
             os.path.join(constants.EXPORT_DIR, src_path)
 
+      # On import force_variant must be True, because if we forced it at
+      # initial install, our only chance when importing it back is that it
+      # works again!
+      self.op.force_variant = True
+
     else: # INSTANCE_CREATE
       if getattr(self.op, "os_type", None) is None:
-        raise errors.OpPrereqError("No guest OS specified")
+        raise errors.OpPrereqError("No guest OS specified",
+                                   errors.ECODE_INVAL)
+      self.op.force_variant = getattr(self.op, "force_variant", False)
 
   def _RunAllocator(self):
     """Run the allocator based on input opcode.
@@ -5516,17 +5906,18 @@ class LUCreateInstance(LogicalUnit):
 
     if not ial.success:
       raise errors.OpPrereqError("Can't compute nodes using"
-                                 " iallocator '%s': %s" % (self.op.iallocator,
-                                                           ial.info))
+                                 " iallocator '%s': %s" %
+                                 (self.op.iallocator, ial.info),
+                                 errors.ECODE_NORES)
     if len(ial.nodes) != ial.required_nodes:
       raise errors.OpPrereqError("iallocator '%s' returned invalid number"
                                  " of nodes (%s), required %s" %
                                  (self.op.iallocator, len(ial.nodes),
-                                  ial.required_nodes))
+                                  ial.required_nodes), errors.ECODE_FAULT)
     self.op.pnode = ial.nodes[0]
     self.LogInfo("Selected nodes for instance %s via iallocator %s: %s",
                  self.op.instance_name, self.op.iallocator,
-                 ", ".join(ial.nodes))
+                 utils.CommaJoin(ial.nodes))
     if ial.required_nodes == 2:
       self.op.snode = ial.nodes[1]
 
@@ -5572,7 +5963,7 @@ class LUCreateInstance(LogicalUnit):
     if (not self.cfg.GetVGName() and
         self.op.disk_template not in constants.DTS_NOT_LVM):
       raise errors.OpPrereqError("Cluster does not support lvm-based"
-                                 " instances")
+                                 " instances", errors.ECODE_STATE)
 
     if self.op.mode == constants.INSTANCE_IMPORT:
       src_node = self.op.src_node
@@ -5593,7 +5984,7 @@ class LUCreateInstance(LogicalUnit):
             break
         if not found:
           raise errors.OpPrereqError("No export found for relative path %s" %
-                                      src_path)
+                                      src_path, errors.ECODE_INVAL)
 
       _CheckNodeOnline(self, src_node)
       result = self.rpc.call_export_info(src_node, src_path)
@@ -5601,12 +5992,14 @@ class LUCreateInstance(LogicalUnit):
 
       export_info = objects.SerializableConfigParser.Loads(str(result.payload))
       if not export_info.has_section(constants.INISECT_EXP):
-        raise errors.ProgrammerError("Corrupted export config")
+        raise errors.ProgrammerError("Corrupted export config",
+                                     errors.ECODE_ENVIRON)
 
       ei_version = export_info.get(constants.INISECT_EXP, 'version')
       if (int(ei_version) != constants.EXPORT_VERSION):
         raise errors.OpPrereqError("Wrong export version %s (wanted %d)" %
-                                   (ei_version, constants.EXPORT_VERSION))
+                                   (ei_version, constants.EXPORT_VERSION),
+                                   errors.ECODE_ENVIRON)
 
       # Check that the new instance doesn't have less disks than the export
       instance_disks = len(self.disks)
@@ -5614,7 +6007,8 @@ class LUCreateInstance(LogicalUnit):
       if instance_disks < export_disks:
         raise errors.OpPrereqError("Not enough disks to import."
                                    " (instance: %d, export: %d)" %
-                                   (instance_disks, export_disks))
+                                   (instance_disks, export_disks),
+                                   errors.ECODE_INVAL)
 
       self.op.os_type = export_info.get(constants.INISECT_EXP, 'os')
       disk_images = []
@@ -5640,15 +6034,13 @@ class LUCreateInstance(LogicalUnit):
             nic.mac = export_info.get(constants.INISECT_INS, nic_mac_ini)
 
     # ENDIF: self.op.mode == constants.INSTANCE_IMPORT
-    # ip ping checks (we use the same ip that was resolved in ExpandNames)
-    if self.op.start and not self.op.ip_check:
-      raise errors.OpPrereqError("Cannot ignore IP address conflicts when"
-                                 " adding an instance in start mode")
 
+    # ip ping checks (we use the same ip that was resolved in ExpandNames)
     if self.op.ip_check:
       if utils.TcpPing(self.check_ip, constants.DEFAULT_NODED_PORT):
         raise errors.OpPrereqError("IP %s of instance %s already in use" %
-                                   (self.check_ip, self.op.instance_name))
+                                   (self.check_ip, self.op.instance_name),
+                                   errors.ECODE_NOTUNIQUE)
 
     #### mac address generation
     # By generating here the mac address both the allocator and the hooks get
@@ -5660,7 +6052,7 @@ class LUCreateInstance(LogicalUnit):
     # creation job will fail.
     for nic in self.nics:
       if nic.mac in (constants.VALUE_AUTO, constants.VALUE_GENERATE):
-        nic.mac = self.cfg.GenerateMAC()
+        nic.mac = self.cfg.GenerateMAC(self.proc.GetECId())
 
     #### allocator run
 
@@ -5675,10 +6067,10 @@ class LUCreateInstance(LogicalUnit):
       "Cannot retrieve locked node %s" % self.op.pnode
     if pnode.offline:
       raise errors.OpPrereqError("Cannot use offline primary node '%s'" %
-                                 pnode.name)
+                                 pnode.name, errors.ECODE_STATE)
     if pnode.drained:
       raise errors.OpPrereqError("Cannot use drained primary node '%s'" %
-                                 pnode.name)
+                                 pnode.name, errors.ECODE_STATE)
 
     self.secondaries = []
 
@@ -5686,10 +6078,10 @@ class LUCreateInstance(LogicalUnit):
     if self.op.disk_template in constants.DTS_NET_MIRROR:
       if self.op.snode is None:
         raise errors.OpPrereqError("The networked disk templates need"
-                                   " a mirror node")
+                                   " a mirror node", errors.ECODE_INVAL)
       if self.op.snode == pnode.name:
-        raise errors.OpPrereqError("The secondary node cannot be"
-                                   " the primary node.")
+        raise errors.OpPrereqError("The secondary node cannot be the"
+                                   " primary node.", errors.ECODE_INVAL)
       _CheckNodeOnline(self, self.op.snode)
       _CheckNodeNotDrained(self, self.op.snode)
       self.secondaries.append(self.op.snode)
@@ -5710,18 +6102,22 @@ class LUCreateInstance(LogicalUnit):
         vg_free = info.get('vg_free', None)
         if not isinstance(vg_free, int):
           raise errors.OpPrereqError("Can't compute free disk space on"
-                                     " node %s" % node)
+                                     " node %s" % node, errors.ECODE_ENVIRON)
         if req_size > vg_free:
           raise errors.OpPrereqError("Not enough disk space on target node %s."
                                      " %d MB available, %d MB required" %
-                                     (node, vg_free, req_size))
+                                     (node, vg_free, req_size),
+                                     errors.ECODE_NORES)
 
     _CheckHVParams(self, nodenames, self.op.hypervisor, self.op.hvparams)
 
     # os verification
     result = self.rpc.call_os_get(pnode.name, self.op.os_type)
     result.Raise("OS '%s' not in supported os list for primary node %s" %
-                 (self.op.os_type, pnode.name), prereq=True)
+                 (self.op.os_type, pnode.name),
+                 prereq=True, ecode=errors.ECODE_INVAL)
+    if not self.op.force_variant:
+      _CheckOSVariant(result.payload, self.op.os_type)
 
     _CheckNicsBridgesExist(self, self.nics, self.pnode.name)
 
@@ -5795,7 +6191,8 @@ class LUCreateInstance(LogicalUnit):
 
     feedback_fn("adding instance %s to cluster config" % instance)
 
-    self.cfg.AddInstance(iobj)
+    self.cfg.AddInstance(iobj, self.proc.GetECId())
+
     # Declare that we don't want to remove the instance lock anymore, as we've
     # added the instance to the config
     del self.remove_locks[locking.LEVEL_INSTANCE]
@@ -5834,7 +6231,8 @@ class LUCreateInstance(LogicalUnit):
     if iobj.disk_template != constants.DT_DISKLESS:
       if self.op.mode == constants.INSTANCE_CREATE:
         feedback_fn("* running the instance OS create scripts...")
-        result = self.rpc.call_instance_os_add(pnode_name, iobj, False)
+        # FIXME: pass debug option from opcode to backend
+        result = self.rpc.call_instance_os_add(pnode_name, iobj, False, 0)
         result.Raise("Could not add os for instance %s"
                      " on node %s" % (instance, pnode_name))
 
@@ -5843,9 +6241,10 @@ class LUCreateInstance(LogicalUnit):
         src_node = self.op.src_node
         src_images = self.src_images
         cluster_name = self.cfg.GetClusterName()
+        # FIXME: pass debug option from opcode to backend
         import_result = self.rpc.call_instance_os_import(pnode_name, iobj,
                                                          src_node, src_images,
-                                                         cluster_name)
+                                                         cluster_name, 0)
         msg = import_result.fail_msg
         if msg:
           self.LogWarning("Error while importing the disk images for instance"
@@ -5857,7 +6256,7 @@ class LUCreateInstance(LogicalUnit):
 
     if self.op.start:
       iobj.admin_up = True
-      self.cfg.Update(iobj)
+      self.cfg.Update(iobj, feedback_fn)
       logging.info("Starting instance %s on node %s", instance, pnode_name)
       feedback_fn("* starting instance...")
       result = self.rpc.call_instance_start(pnode_name, iobj, None, None)
@@ -5947,7 +6346,7 @@ class LUReplaceDisks(LogicalUnit):
       remote_node = self.cfg.ExpandNodeName(self.op.remote_node)
       if remote_node is None:
         raise errors.OpPrereqError("Node '%s' not known" %
-                                   self.op.remote_node)
+                                   self.op.remote_node, errors.ECODE_NOENT)
 
       self.op.remote_node = remote_node
 
@@ -5964,7 +6363,7 @@ class LUReplaceDisks(LogicalUnit):
 
     self.replacer = TLReplaceDisks(self, self.op.instance_name, self.op.mode,
                                    self.op.iallocator, self.op.remote_node,
-                                   self.op.disks)
+                                   self.op.disks, False)
 
     self.tasklets = [self.replacer]
 
@@ -6019,7 +6418,8 @@ class LUEvacuateNode(LogicalUnit):
   def ExpandNames(self):
     self.op.node_name = self.cfg.ExpandNodeName(self.op.node_name)
     if self.op.node_name is None:
-      raise errors.OpPrereqError("Node '%s' not known" % self.op.node_name)
+      raise errors.OpPrereqError("Node '%s' not known" % self.op.node_name,
+                                 errors.ECODE_NOENT)
 
     self.needed_locks = {}
 
@@ -6031,7 +6431,7 @@ class LUEvacuateNode(LogicalUnit):
       remote_node = self.cfg.ExpandNodeName(self.op.remote_node)
       if remote_node is None:
         raise errors.OpPrereqError("Node '%s' not known" %
-                                   self.op.remote_node)
+                                   self.op.remote_node, errors.ECODE_NOENT)
 
       self.op.remote_node = remote_node
 
@@ -6043,7 +6443,7 @@ class LUEvacuateNode(LogicalUnit):
       self.recalculate_locks[locking.LEVEL_NODE] = constants.LOCKS_APPEND
 
     else:
-      raise errors.OpPrereqError("Invalid parameters")
+      raise errors.OpPrereqError("Invalid parameters", errors.ECODE_INVAL)
 
     # Create tasklets for replacing disks for all secondary instances on this
     # node
@@ -6055,7 +6455,8 @@ class LUEvacuateNode(LogicalUnit):
       names.append(inst.name)
 
       replacer = TLReplaceDisks(self, inst.name, constants.REPLACE_DISK_CHG,
-                                self.op.iallocator, self.op.remote_node, [])
+                                self.op.iallocator, self.op.remote_node, [],
+                                True)
       tasklets.append(replacer)
 
     self.tasklets = tasklets
@@ -6097,7 +6498,7 @@ class TLReplaceDisks(Tasklet):
 
   """
   def __init__(self, lu, instance_name, mode, iallocator_name, remote_node,
-               disks):
+               disks, delay_iallocator):
     """Initializes this class.
 
     """
@@ -6109,6 +6510,7 @@ class TLReplaceDisks(Tasklet):
     self.iallocator_name = iallocator_name
     self.remote_node = remote_node
     self.disks = disks
+    self.delay_iallocator = delay_iallocator
 
     # Runtime data
     self.instance = None
@@ -6128,17 +6530,17 @@ class TLReplaceDisks(Tasklet):
       if remote_node is None and iallocator is None:
         raise errors.OpPrereqError("When changing the secondary either an"
                                    " iallocator script must be used or the"
-                                   " new node given")
+                                   " new node given", errors.ECODE_INVAL)
 
       if remote_node is not None and iallocator is not None:
         raise errors.OpPrereqError("Give either the iallocator or the new"
-                                   " secondary, not both")
+                                   " secondary, not both", errors.ECODE_INVAL)
 
     elif remote_node is not None or iallocator is not None:
       # Not replacing the secondary
       raise errors.OpPrereqError("The iallocator and new node options can"
                                  " only be used when changing the"
-                                 " secondary node")
+                                 " secondary node", errors.ECODE_INVAL)
 
   @staticmethod
   def _RunAllocator(lu, iallocator_name, instance_name, relocate_from):
@@ -6154,12 +6556,15 @@ class TLReplaceDisks(Tasklet):
 
     if not ial.success:
       raise errors.OpPrereqError("Can't compute nodes using iallocator '%s':"
-                                 " %s" % (iallocator_name, ial.info))
+                                 " %s" % (iallocator_name, ial.info),
+                                 errors.ECODE_NORES)
 
     if len(ial.nodes) != ial.required_nodes:
       raise errors.OpPrereqError("iallocator '%s' returned invalid number"
                                  " of nodes (%s), required %s" %
-                                 (len(ial.nodes), ial.required_nodes))
+                                 (iallocator_name,
+                                  len(ial.nodes), ial.required_nodes),
+                                 errors.ECODE_FAULT)
 
     remote_node_name = ial.nodes[0]
 
@@ -6178,26 +6583,40 @@ class TLReplaceDisks(Tasklet):
     This checks that the instance is in the cluster.
 
     """
-    self.instance = self.cfg.GetInstanceInfo(self.instance_name)
-    assert self.instance is not None, \
+    self.instance = instance = self.cfg.GetInstanceInfo(self.instance_name)
+    assert instance is not None, \
       "Cannot retrieve locked instance %s" % self.instance_name
 
-    if self.instance.disk_template != constants.DT_DRBD8:
+    if instance.disk_template != constants.DT_DRBD8:
       raise errors.OpPrereqError("Can only run replace disks for DRBD8-based"
-                                 " instances")
+                                 " instances", errors.ECODE_INVAL)
 
-    if len(self.instance.secondary_nodes) != 1:
+    if len(instance.secondary_nodes) != 1:
       raise errors.OpPrereqError("The instance has a strange layout,"
                                  " expected one secondary but found %d" %
-                                 len(self.instance.secondary_nodes))
+                                 len(instance.secondary_nodes),
+                                 errors.ECODE_FAULT)
+
+    if not self.delay_iallocator:
+      self._CheckPrereq2()
+
+  def _CheckPrereq2(self):
+    """Check prerequisites, second part.
+
+    This function should always be part of CheckPrereq. It was separated and is
+    now called from Exec because during node evacuation iallocator was only
+    called with an unmodified cluster model, not taking planned changes into
+    account.
 
-    secondary_node = self.instance.secondary_nodes[0]
+    """
+    instance = self.instance
+    secondary_node = instance.secondary_nodes[0]
 
     if self.iallocator_name is None:
       remote_node = self.remote_node
     else:
       remote_node = self._RunAllocator(self.lu, self.iallocator_name,
-                                       self.instance.name, secondary_node)
+                                       instance.name, instance.secondary_nodes)
 
     if remote_node is not None:
       self.remote_node_info = self.cfg.GetNodeInfo(remote_node)
@@ -6208,34 +6627,37 @@ class TLReplaceDisks(Tasklet):
 
     if remote_node == self.instance.primary_node:
       raise errors.OpPrereqError("The specified node is the primary node of"
-                                 " the instance.")
+                                 " the instance.", errors.ECODE_INVAL)
 
     if remote_node == secondary_node:
       raise errors.OpPrereqError("The specified node is already the"
-                                 " secondary node of the instance.")
+                                 " secondary node of the instance.",
+                                 errors.ECODE_INVAL)
 
     if self.disks and self.mode in (constants.REPLACE_DISK_AUTO,
                                     constants.REPLACE_DISK_CHG):
-      raise errors.OpPrereqError("Cannot specify disks to be replaced")
+      raise errors.OpPrereqError("Cannot specify disks to be replaced",
+                                 errors.ECODE_INVAL)
 
     if self.mode == constants.REPLACE_DISK_AUTO:
-      faulty_primary = self._FindFaultyDisks(self.instance.primary_node)
+      faulty_primary = self._FindFaultyDisks(instance.primary_node)
       faulty_secondary = self._FindFaultyDisks(secondary_node)
 
       if faulty_primary and faulty_secondary:
         raise errors.OpPrereqError("Instance %s has faulty disks on more than"
                                    " one node and can not be repaired"
-                                   " automatically" % self.instance_name)
+                                   " automatically" % self.instance_name,
+                                   errors.ECODE_STATE)
 
       if faulty_primary:
         self.disks = faulty_primary
-        self.target_node = self.instance.primary_node
+        self.target_node = instance.primary_node
         self.other_node = secondary_node
         check_nodes = [self.target_node, self.other_node]
       elif faulty_secondary:
         self.disks = faulty_secondary
         self.target_node = secondary_node
-        self.other_node = self.instance.primary_node
+        self.other_node = instance.primary_node
         check_nodes = [self.target_node, self.other_node]
       else:
         self.disks = []
@@ -6244,18 +6666,18 @@ class TLReplaceDisks(Tasklet):
     else:
       # Non-automatic modes
       if self.mode == constants.REPLACE_DISK_PRI:
-        self.target_node = self.instance.primary_node
+        self.target_node = instance.primary_node
         self.other_node = secondary_node
         check_nodes = [self.target_node, self.other_node]
 
       elif self.mode == constants.REPLACE_DISK_SEC:
         self.target_node = secondary_node
-        self.other_node = self.instance.primary_node
+        self.other_node = instance.primary_node
         check_nodes = [self.target_node, self.other_node]
 
       elif self.mode == constants.REPLACE_DISK_CHG:
         self.new_node = remote_node
-        self.other_node = self.instance.primary_node
+        self.other_node = instance.primary_node
         self.target_node = secondary_node
         check_nodes = [self.new_node, self.other_node]
 
@@ -6274,7 +6696,7 @@ class TLReplaceDisks(Tasklet):
 
     # Check whether disks are valid
     for disk_idx in self.disks:
-      self.instance.FindDisk(disk_idx)
+      instance.FindDisk(disk_idx)
 
     # Get secondary node IP addresses
     node_2nd_ip = {}
@@ -6291,12 +6713,15 @@ class TLReplaceDisks(Tasklet):
     This dispatches the disk replacement to the appropriate handler.
 
     """
+    if self.delay_iallocator:
+      self._CheckPrereq2()
+
     if not self.disks:
       feedback_fn("No disks need replacement")
       return
 
     feedback_fn("Replacing disk(s) %s for %s" %
-                (", ".join([str(i) for i in self.disks]), self.instance.name))
+                (utils.CommaJoin(self.disks), self.instance.name))
 
     activate_disks = (not self.instance.admin_up)
 
@@ -6307,12 +6732,15 @@ class TLReplaceDisks(Tasklet):
     try:
       # Should we replace the secondary node?
       if self.new_node is not None:
-        return self._ExecDrbd8Secondary()
+        fn = self._ExecDrbd8Secondary
       else:
-        return self._ExecDrbd8DiskOnly()
+        fn = self._ExecDrbd8DiskOnly
+
+      return fn(feedback_fn)
 
     finally:
-      # Deactivate the instance disks if we're replacing them on a down instance
+      # Deactivate the instance disks if we're replacing them on a
+      # down instance
       if activate_disks:
         _SafeShutdownInstanceDisks(self.lu, self.instance)
 
@@ -6398,7 +6826,7 @@ class TLReplaceDisks(Tasklet):
     return iv_names
 
   def _CheckDevices(self, node_name, iv_names):
-    for name, (dev, old_lvs, new_lvs) in iv_names.iteritems():
+    for name, (dev, _, _) in iv_names.iteritems():
       self.cfg.SetDiskID(dev, node_name)
 
       result = self.rpc.call_blockdev_find(node_name, dev)
@@ -6414,7 +6842,7 @@ class TLReplaceDisks(Tasklet):
         raise errors.OpExecError("DRBD device %s is degraded!" % name)
 
   def _RemoveOldStorage(self, node_name, iv_names):
-    for name, (dev, old_lvs, _) in iv_names.iteritems():
+    for name, (_, old_lvs, _) in iv_names.iteritems():
       self.lu.LogInfo("Remove logical volumes for %s" % name)
 
       for lv in old_lvs:
@@ -6425,7 +6853,7 @@ class TLReplaceDisks(Tasklet):
           self.lu.LogWarning("Can't remove old LV: %s" % msg,
                              hint="remove unused LVs manually")
 
-  def _ExecDrbd8DiskOnly(self):
+  def _ExecDrbd8DiskOnly(self, feedback_fn):
     """Replace a disk on the primary or secondary for DRBD 8.
 
     The algorithm for replace is quite complicated:
@@ -6533,13 +6961,13 @@ class TLReplaceDisks(Tasklet):
 
       dev.children = new_lvs
 
-      self.cfg.Update(self.instance)
+      self.cfg.Update(self.instance, feedback_fn)
 
     # Wait for sync
     # This can fail as the old devices are degraded and _WaitForSync
     # does a combined result over all disks, so we don't check its return value
     self.lu.LogStep(5, steps_total, "Sync devices")
-    _WaitForSync(self.lu, self.instance, unlock=True)
+    _WaitForSync(self.lu, self.instance)
 
     # Check all devices manually
     self._CheckDevices(self.instance.primary_node, iv_names)
@@ -6548,7 +6976,7 @@ class TLReplaceDisks(Tasklet):
     self.lu.LogStep(6, steps_total, "Removing old storage")
     self._RemoveOldStorage(self.target_node, iv_names)
 
-  def _ExecDrbd8Secondary(self):
+  def _ExecDrbd8Secondary(self, feedback_fn):
     """Replace the secondary node for DRBD 8.
 
     The algorithm for replace is quite complicated:
@@ -6595,7 +7023,7 @@ class TLReplaceDisks(Tasklet):
     minors = self.cfg.AllocateDRBDMinor([self.new_node
                                          for dev in self.instance.disks],
                                         self.instance.name)
-    logging.debug("Allocated minors %r" % (minors,))
+    logging.debug("Allocated minors %r", minors)
 
     iv_names = {}
     for idx, (dev, new_minor) in enumerate(zip(self.instance.disks, minors)):
@@ -6609,6 +7037,7 @@ class TLReplaceDisks(Tasklet):
       if self.instance.primary_node == o_node1:
         p_minor = o_minor1
       else:
+        assert self.instance.primary_node == o_node2, "Three-node instance?"
         p_minor = o_minor2
 
       new_alone_id = (self.instance.primary_node, self.new_node, None,
@@ -6661,7 +7090,7 @@ class TLReplaceDisks(Tasklet):
       dev.logical_id = new_logical_id
       self.cfg.SetDiskID(dev, self.instance.primary_node)
 
-    self.cfg.Update(self.instance)
+    self.cfg.Update(self.instance, feedback_fn)
 
     # and now perform the drbd attach
     self.lu.LogInfo("Attaching primary drbds to new secondary"
@@ -6684,7 +7113,7 @@ class TLReplaceDisks(Tasklet):
     # This can fail as the old devices are degraded and _WaitForSync
     # does a combined result over all disks, so we don't check its return value
     self.lu.LogStep(5, steps_total, "Sync devices")
-    _WaitForSync(self.lu, self.instance, unlock=True)
+    _WaitForSync(self.lu, self.instance)
 
     # Check all devices manually
     self._CheckDevices(self.instance.primary_node, iv_names)
@@ -6704,7 +7133,8 @@ class LURepairNodeStorage(NoHooksLU):
   def CheckArguments(self):
     node_name = self.cfg.ExpandNodeName(self.op.node_name)
     if node_name is None:
-      raise errors.OpPrereqError("Invalid node name '%s'" % self.op.node_name)
+      raise errors.OpPrereqError("Invalid node name '%s'" % self.op.node_name,
+                                 errors.ECODE_NOENT)
 
     self.op.node_name = node_name
 
@@ -6714,10 +7144,18 @@ class LURepairNodeStorage(NoHooksLU):
       }
 
   def _CheckFaultyDisks(self, instance, node_name):
-    if _FindFaultyInstanceDisks(self.cfg, self.rpc, instance,
-                                node_name, True):
-      raise errors.OpPrereqError("Instance '%s' has faulty disks on"
-                                 " node '%s'" % (instance.name, node_name))
+    """Ensure faulty disks abort the opcode or at least warn."""
+    try:
+      if _FindFaultyInstanceDisks(self.cfg, self.rpc, instance,
+                                  node_name, True):
+        raise errors.OpPrereqError("Instance '%s' has faulty disks on"
+                                   " node '%s'" % (instance.name, node_name),
+                                   errors.ECODE_STATE)
+    except errors.OpPrereqError, err:
+      if self.op.ignore_consistency:
+        self.proc.LogWarning(str(err.args[0]))
+      else:
+        raise
 
   def CheckPrereq(self):
     """Check prerequisites.
@@ -6728,10 +7166,13 @@ class LURepairNodeStorage(NoHooksLU):
     if (constants.SO_FIX_CONSISTENCY not in
         constants.VALID_STORAGE_OPERATIONS.get(storage_type, [])):
       raise errors.OpPrereqError("Storage units of type '%s' can not be"
-                                 " repaired" % storage_type)
+                                 " repaired" % storage_type,
+                                 errors.ECODE_INVAL)
 
     # Check whether any instance on this node has faulty disks
     for inst in _GetNodeInstances(self.cfg, self.op.node_name):
+      if not inst.admin_up:
+        continue
       check_nodes = set(inst.all_nodes)
       check_nodes.discard(self.op.node_name)
       for inst_node_name in check_nodes:
@@ -6803,7 +7244,7 @@ class LUGrowDisk(LogicalUnit):
 
     if instance.disk_template not in (constants.DT_PLAIN, constants.DT_DRBD8):
       raise errors.OpPrereqError("Instance's disk layout does not support"
-                                 " growing.")
+                                 " growing.", errors.ECODE_INVAL)
 
     self.disk = instance.FindDisk(self.op.disk)
 
@@ -6815,11 +7256,12 @@ class LUGrowDisk(LogicalUnit):
       vg_free = info.payload.get('vg_free', None)
       if not isinstance(vg_free, int):
         raise errors.OpPrereqError("Can't compute free disk space on"
-                                   " node %s" % node)
+                                   " node %s" % node, errors.ECODE_ENVIRON)
       if self.op.amount > vg_free:
         raise errors.OpPrereqError("Not enough disk space on target node %s:"
                                    " %d MiB available, %d MiB required" %
-                                   (node, vg_free, self.op.amount))
+                                   (node, vg_free, self.op.amount),
+                                   errors.ECODE_NORES)
 
   def Exec(self, feedback_fn):
     """Execute disk grow.
@@ -6831,8 +7273,16 @@ class LUGrowDisk(LogicalUnit):
       self.cfg.SetDiskID(disk, node)
       result = self.rpc.call_blockdev_grow(node, disk, self.op.amount)
       result.Raise("Grow request failed to node %s" % node)
+
+      # TODO: Rewrite code to work properly
+      # DRBD goes into sync mode for a short amount of time after executing the
+      # "resize" command. DRBD 8.x below version 8.0.13 contains a bug whereby
+      # calling "resize" in sync mode fails. Sleeping for a short amount of
+      # time is a work-around.
+      time.sleep(5)
+
     disk.RecordGrow(self.op.amount)
-    self.cfg.Update(instance)
+    self.cfg.Update(instance, feedback_fn)
     if self.op.wait_for_sync:
       disk_abort = not _WaitForSync(self, instance)
       if disk_abort:
@@ -6852,14 +7302,16 @@ class LUQueryInstanceData(NoHooksLU):
     self.share_locks = dict.fromkeys(locking.LEVELS, 1)
 
     if not isinstance(self.op.instances, list):
-      raise errors.OpPrereqError("Invalid argument type 'instances'")
+      raise errors.OpPrereqError("Invalid argument type 'instances'",
+                                 errors.ECODE_INVAL)
 
     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)
+          raise errors.OpPrereqError("Instance '%s' not known" % name,
+                                     errors.ECODE_NOENT)
         self.wanted_names.append(full_name)
       self.needed_locks[locking.LEVEL_INSTANCE] = self.wanted_names
     else:
@@ -6984,12 +7436,13 @@ class LUQueryInstanceData(NoHooksLU):
         "hypervisor": instance.hypervisor,
         "network_port": instance.network_port,
         "hv_instance": instance.hvparams,
-        "hv_actual": cluster.FillHV(instance),
+        "hv_actual": cluster.FillHV(instance, skip_globals=True),
         "be_instance": instance.beparams,
         "be_actual": cluster.FillBE(instance),
         "serial_no": instance.serial_no,
         "mtime": instance.mtime,
         "ctime": instance.ctime,
+        "uuid": instance.uuid,
         }
 
       result[instance.name] = idict
@@ -7018,7 +7471,10 @@ class LUSetInstanceParams(LogicalUnit):
     self.op.force = getattr(self.op, "force", False)
     if not (self.op.nics or self.op.disks or
             self.op.hvparams or self.op.beparams):
-      raise errors.OpPrereqError("No changes submitted")
+      raise errors.OpPrereqError("No changes submitted", errors.ECODE_INVAL)
+
+    if self.op.hvparams:
+      _CheckGlobalHvParams(self.op.hvparams)
 
     # Disk validation
     disk_addremove = 0
@@ -7030,33 +7486,35 @@ class LUSetInstanceParams(LogicalUnit):
         disk_addremove += 1
       else:
         if not isinstance(disk_op, int):
-          raise errors.OpPrereqError("Invalid disk index")
+          raise errors.OpPrereqError("Invalid disk index", errors.ECODE_INVAL)
         if not isinstance(disk_dict, dict):
           msg = "Invalid disk value: expected dict, got '%s'" % disk_dict
-          raise errors.OpPrereqError(msg)
+          raise errors.OpPrereqError(msg, errors.ECODE_INVAL)
 
       if disk_op == constants.DDM_ADD:
         mode = disk_dict.setdefault('mode', constants.DISK_RDWR)
         if mode not in constants.DISK_ACCESS_SET:
-          raise errors.OpPrereqError("Invalid disk access mode '%s'" % mode)
+          raise errors.OpPrereqError("Invalid disk access mode '%s'" % mode,
+                                     errors.ECODE_INVAL)
         size = disk_dict.get('size', None)
         if size is None:
-          raise errors.OpPrereqError("Required disk parameter size missing")
+          raise errors.OpPrereqError("Required disk parameter size missing",
+                                     errors.ECODE_INVAL)
         try:
           size = int(size)
-        except ValueError, err:
+        except (TypeError, ValueError), err:
           raise errors.OpPrereqError("Invalid disk size parameter: %s" %
-                                     str(err))
+                                     str(err), errors.ECODE_INVAL)
         disk_dict['size'] = size
       else:
         # modification of disk
         if 'size' in disk_dict:
           raise errors.OpPrereqError("Disk size change not possible, use"
-                                     " grow-disk")
+                                     " grow-disk", errors.ECODE_INVAL)
 
     if disk_addremove > 1:
       raise errors.OpPrereqError("Only one disk add or remove operation"
-                                 " supported at a time")
+                                 " supported at a time", errors.ECODE_INVAL)
 
     # NIC validation
     nic_addremove = 0
@@ -7068,10 +7526,10 @@ class LUSetInstanceParams(LogicalUnit):
         nic_addremove += 1
       else:
         if not isinstance(nic_op, int):
-          raise errors.OpPrereqError("Invalid nic index")
+          raise errors.OpPrereqError("Invalid nic index", errors.ECODE_INVAL)
         if not isinstance(nic_dict, dict):
           msg = "Invalid nic value: expected dict, got '%s'" % nic_dict
-          raise errors.OpPrereqError(msg)
+          raise errors.OpPrereqError(msg, errors.ECODE_INVAL)
 
       # nic_dict should be a dict
       nic_ip = nic_dict.get('ip', None)
@@ -7080,13 +7538,14 @@ class LUSetInstanceParams(LogicalUnit):
           nic_dict['ip'] = None
         else:
           if not utils.IsValidIP(nic_ip):
-            raise errors.OpPrereqError("Invalid IP address '%s'" % nic_ip)
+            raise errors.OpPrereqError("Invalid IP address '%s'" % nic_ip,
+                                       errors.ECODE_INVAL)
 
       nic_bridge = nic_dict.get('bridge', None)
       nic_link = nic_dict.get('link', None)
       if nic_bridge and nic_link:
         raise errors.OpPrereqError("Cannot pass 'bridge' and 'link'"
-                                   " at the same time")
+                                   " at the same time", errors.ECODE_INVAL)
       elif nic_bridge and nic_bridge.lower() == constants.VALUE_NONE:
         nic_dict['bridge'] = None
       elif nic_link and nic_link.lower() == constants.VALUE_NONE:
@@ -7100,15 +7559,16 @@ class LUSetInstanceParams(LogicalUnit):
       if 'mac' in nic_dict:
         nic_mac = nic_dict['mac']
         if nic_mac not in (constants.VALUE_AUTO, constants.VALUE_GENERATE):
-          if not utils.IsValidMac(nic_mac):
-            raise errors.OpPrereqError("Invalid MAC address %s" % nic_mac)
+          nic_mac = utils.NormalizeAndValidateMac(nic_mac)
+
         if nic_op != constants.DDM_ADD and nic_mac == constants.VALUE_AUTO:
           raise errors.OpPrereqError("'auto' is not a valid MAC address when"
-                                     " modifying an existing nic")
+                                     " modifying an existing nic",
+                                     errors.ECODE_INVAL)
 
     if nic_addremove > 1:
       raise errors.OpPrereqError("Only one NIC add or remove operation"
-                                 " supported at a time")
+                                 " supported at a time", errors.ECODE_INVAL)
 
   def ExpandNames(self):
     self._ExpandAndLockInstance()
@@ -7170,7 +7630,8 @@ class LUSetInstanceParams(LogicalUnit):
     nl = [self.cfg.GetMasterNode()] + list(self.instance.all_nodes)
     return env, nl, nl
 
-  def _GetUpdatedParams(self, old_params, update_dict,
+  @staticmethod
+  def _GetUpdatedParams(old_params, update_dict,
                         default_values, parameter_types):
     """Return the new params dict for the given params.
 
@@ -7281,7 +7742,8 @@ class LUSetInstanceParams(LogicalUnit):
         if miss_mem > 0:
           raise errors.OpPrereqError("This change will prevent the instance"
                                      " from starting, due to %d MB of memory"
-                                     " missing on its primary node" % miss_mem)
+                                     " missing on its primary node" % miss_mem,
+                                     errors.ECODE_NORES)
 
       if be_new[constants.BE_AUTO_BALANCE]:
         for node, nres in nodeinfo.items():
@@ -7304,14 +7766,20 @@ class LUSetInstanceParams(LogicalUnit):
     for nic_op, nic_dict in self.op.nics:
       if nic_op == constants.DDM_REMOVE:
         if not instance.nics:
-          raise errors.OpPrereqError("Instance has no NICs, cannot remove")
+          raise errors.OpPrereqError("Instance has no NICs, cannot remove",
+                                     errors.ECODE_INVAL)
         continue
       if nic_op != constants.DDM_ADD:
         # an existing nic
+        if not instance.nics:
+          raise errors.OpPrereqError("Invalid NIC index %s, instance has"
+                                     " no NICs" % nic_op,
+                                     errors.ECODE_INVAL)
         if nic_op < 0 or nic_op >= len(instance.nics):
           raise errors.OpPrereqError("Invalid NIC index %s, valid values"
                                      " are 0 to %d" %
-                                     (nic_op, len(instance.nics)))
+                                     (nic_op, len(instance.nics) - 1),
+                                     errors.ECODE_INVAL)
         old_nic_params = instance.nics[nic_op].nicparams
         old_nic_ip = instance.nics[nic_op].ip
       else:
@@ -7342,7 +7810,7 @@ class LUSetInstanceParams(LogicalUnit):
           if self.force:
             self.warn.append(msg)
           else:
-            raise errors.OpPrereqError(msg)
+            raise errors.OpPrereqError(msg, errors.ECODE_ENVIRON)
       if new_nic_mode == constants.NIC_MODE_ROUTED:
         if 'ip' in nic_dict:
           nic_ip = nic_dict['ip']
@@ -7350,49 +7818,57 @@ class LUSetInstanceParams(LogicalUnit):
           nic_ip = old_nic_ip
         if nic_ip is None:
           raise errors.OpPrereqError('Cannot set the nic ip to None'
-                                     ' on a routed nic')
+                                     ' on a routed nic', errors.ECODE_INVAL)
       if 'mac' in nic_dict:
         nic_mac = nic_dict['mac']
         if nic_mac is None:
-          raise errors.OpPrereqError('Cannot set the nic mac to None')
+          raise errors.OpPrereqError('Cannot set the nic mac to None',
+                                     errors.ECODE_INVAL)
         elif nic_mac in (constants.VALUE_AUTO, constants.VALUE_GENERATE):
           # otherwise generate the mac
-          nic_dict['mac'] = self.cfg.GenerateMAC()
+          nic_dict['mac'] = self.cfg.GenerateMAC(self.proc.GetECId())
         else:
           # or validate/reserve the current one
-          if self.cfg.IsMacInUse(nic_mac):
+          try:
+            self.cfg.ReserveMAC(nic_mac, self.proc.GetECId())
+          except errors.ReservationError:
             raise errors.OpPrereqError("MAC address %s already in use"
-                                       " in cluster" % nic_mac)
+                                       " in cluster" % nic_mac,
+                                       errors.ECODE_NOTUNIQUE)
 
     # DISK processing
     if self.op.disks and instance.disk_template == constants.DT_DISKLESS:
       raise errors.OpPrereqError("Disk operations not supported for"
-                                 " diskless instances")
-    for disk_op, disk_dict in self.op.disks:
+                                 " diskless instances",
+                                 errors.ECODE_INVAL)
+    for disk_op, _ in self.op.disks:
       if disk_op == constants.DDM_REMOVE:
         if len(instance.disks) == 1:
           raise errors.OpPrereqError("Cannot remove the last disk of"
-                                     " an instance")
+                                     " an instance",
+                                     errors.ECODE_INVAL)
         ins_l = self.rpc.call_instance_list([pnode], [instance.hypervisor])
         ins_l = ins_l[pnode]
         msg = ins_l.fail_msg
         if msg:
           raise errors.OpPrereqError("Can't contact node %s: %s" %
-                                     (pnode, msg))
+                                     (pnode, msg), errors.ECODE_ENVIRON)
         if instance.name in ins_l.payload:
           raise errors.OpPrereqError("Instance is running, can't remove"
-                                     " disks.")
+                                     " disks.", errors.ECODE_STATE)
 
       if (disk_op == constants.DDM_ADD and
           len(instance.nics) >= constants.MAX_DISKS):
         raise errors.OpPrereqError("Instance has too many disks (%d), cannot"
-                                   " add more" % constants.MAX_DISKS)
+                                   " add more" % constants.MAX_DISKS,
+                                   errors.ECODE_STATE)
       if disk_op not in (constants.DDM_ADD, constants.DDM_REMOVE):
         # an existing disk
         if disk_op < 0 or disk_op >= len(instance.disks):
           raise errors.OpPrereqError("Invalid disk index %s, valid values"
                                      " are 0 to %d" %
-                                     (disk_op, len(instance.disks)))
+                                     (disk_op, len(instance.disks)),
+                                     errors.ECODE_INVAL)
 
     return
 
@@ -7409,7 +7885,6 @@ class LUSetInstanceParams(LogicalUnit):
 
     result = []
     instance = self.instance
-    cluster = self.cluster
     # disk changes
     for disk_op, disk_dict in self.op.disks:
       if disk_op == constants.DDM_REMOVE:
@@ -7484,8 +7959,8 @@ class LUSetInstanceParams(LogicalUnit):
         for key in 'mac', 'ip':
           if key in nic_dict:
             setattr(instance.nics[nic_op], key, nic_dict[key])
-        if nic_op in self.nic_pnew:
-          instance.nics[nic_op].nicparams = self.nic_pnew[nic_op]
+        if nic_op in self.nic_pinst:
+          instance.nics[nic_op].nicparams = self.nic_pinst[nic_op]
         for key, val in nic_dict.iteritems():
           result.append(("nic.%s/%d" % (key, nic_op), val))
 
@@ -7501,7 +7976,7 @@ class LUSetInstanceParams(LogicalUnit):
       for key, val in self.op.beparams.iteritems():
         result.append(("be/%s" % key, val))
 
-    self.cfg.Update(instance)
+    self.cfg.Update(instance, feedback_fn)
 
     return result
 
@@ -7557,6 +8032,13 @@ 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
@@ -7582,6 +8064,7 @@ 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,
@@ -7605,7 +8088,8 @@ class LUExportInstance(LogicalUnit):
 
     if self.dst_node is None:
       # This is wrong node name, not a non-locked node
-      raise errors.OpPrereqError("Wrong node name %s" % self.op.target_node)
+      raise errors.OpPrereqError("Wrong node name %s" % self.op.target_node,
+                                 errors.ECODE_NOENT)
     _CheckNodeOnline(self, self.dst_node.name)
     _CheckNodeNotDrained(self, self.dst_node.name)
 
@@ -7613,7 +8097,7 @@ class LUExportInstance(LogicalUnit):
     for disk in self.instance.disks:
       if disk.dev_type == constants.LD_FILE:
         raise errors.OpPrereqError("Export not supported for instances with"
-                                   " file-based disks")
+                                   " file-based disks", errors.ECODE_INVAL)
 
   def Exec(self, feedback_fn):
     """Export an instance to an image in the cluster.
@@ -7626,7 +8110,8 @@ 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))
 
@@ -7639,67 +8124,84 @@ class LUExportInstance(LogicalUnit):
     for disk in instance.disks:
       self.cfg.SetDiskID(disk, src_node)
 
-    # per-disk results
-    dresults = []
+    activate_disks = (not instance.admin_up)
+
+    if activate_disks:
+      # Activate the instance disks if we'exporting a stopped instance
+      feedback_fn("Activating disks for %s" % instance.name)
+      _StartInstanceDisks(self, instance, None)
+
     try:
-      for idx, disk in enumerate(instance.disks):
-        feedback_fn("Creating a snapshot of disk/%s on node %s" %
-                    (idx, src_node))
+      # per-disk results
+      dresults = []
+      try:
+        for idx, disk in enumerate(instance.disks):
+          feedback_fn("Creating a snapshot of disk/%s on node %s" %
+                      (idx, src_node))
+
+          # result.payload will be a snapshot of an lvm leaf of the one we
+          # passed
+          result = self.rpc.call_blockdev_snapshot(src_node, disk)
+          msg = result.fail_msg
+          if msg:
+            self.LogWarning("Could not snapshot disk/%s on node %s: %s",
+                            idx, src_node, msg)
+            snap_disks.append(False)
+          else:
+            disk_id = (vgname, result.payload)
+            new_dev = objects.Disk(dev_type=constants.LD_LV, size=disk.size,
+                                   logical_id=disk_id, physical_id=disk_id,
+                                   iv_name=disk.iv_name)
+            snap_disks.append(new_dev)
 
-        # result.payload will be a snapshot of an lvm leaf of the one we passed
-        result = self.rpc.call_blockdev_snapshot(src_node, disk)
-        msg = result.fail_msg
-        if msg:
-          self.LogWarning("Could not snapshot disk/%s on node %s: %s",
-                          idx, src_node, msg)
-          snap_disks.append(False)
+      finally:
+        if self.op.shutdown and instance.admin_up:
+          feedback_fn("Starting instance %s" % instance.name)
+          result = self.rpc.call_instance_start(src_node, instance, None, None)
+          msg = result.fail_msg
+          if msg:
+            _ShutdownInstanceDisks(self, instance)
+            raise errors.OpExecError("Could not start instance: %s" % msg)
+
+      # TODO: check for size
+
+      cluster_name = self.cfg.GetClusterName()
+      for idx, dev in enumerate(snap_disks):
+        feedback_fn("Exporting snapshot %s from %s to %s" %
+                    (idx, src_node, dst_node.name))
+        if dev:
+          # FIXME: pass debug from opcode to backend
+          result = self.rpc.call_snapshot_export(src_node, dev, dst_node.name,
+                                                 instance, cluster_name,
+                                                 idx, 0)
+          msg = result.fail_msg
+          if msg:
+            self.LogWarning("Could not export disk/%s from node %s to"
+                            " node %s: %s", idx, src_node, dst_node.name, msg)
+            dresults.append(False)
+          else:
+            dresults.append(True)
+          msg = self.rpc.call_blockdev_remove(src_node, dev).fail_msg
+          if msg:
+            self.LogWarning("Could not remove snapshot for disk/%d from node"
+                            " %s: %s", idx, src_node, msg)
         else:
-          disk_id = (vgname, result.payload)
-          new_dev = objects.Disk(dev_type=constants.LD_LV, size=disk.size,
-                                 logical_id=disk_id, physical_id=disk_id,
-                                 iv_name=disk.iv_name)
-          snap_disks.append(new_dev)
-
-    finally:
-      if self.op.shutdown and instance.admin_up:
-        feedback_fn("Starting instance %s" % instance.name)
-        result = self.rpc.call_instance_start(src_node, instance, None, None)
-        msg = result.fail_msg
-        if msg:
-          _ShutdownInstanceDisks(self, instance)
-          raise errors.OpExecError("Could not start instance: %s" % msg)
-
-    # TODO: check for size
-
-    cluster_name = self.cfg.GetClusterName()
-    for idx, dev in enumerate(snap_disks):
-      feedback_fn("Exporting snapshot %s from %s to %s" %
-                  (idx, src_node, dst_node.name))
-      if dev:
-        result = self.rpc.call_snapshot_export(src_node, dev, dst_node.name,
-                                               instance, cluster_name, idx)
-        msg = result.fail_msg
-        if msg:
-          self.LogWarning("Could not export disk/%s from node %s to"
-                          " node %s: %s", idx, src_node, dst_node.name, msg)
           dresults.append(False)
-        else:
-          dresults.append(True)
-        msg = self.rpc.call_blockdev_remove(src_node, dev).fail_msg
-        if msg:
-          self.LogWarning("Could not remove snapshot for disk/%d from node"
-                          " %s: %s", idx, src_node, msg)
-      else:
-        dresults.append(False)
 
-    feedback_fn("Finalizing export on %s" % dst_node.name)
-    result = self.rpc.call_finalize_export(dst_node.name, instance, snap_disks)
-    fin_resu = True
-    msg = result.fail_msg
-    if msg:
-      self.LogWarning("Could not finalize export for instance %s"
-                      " on node %s: %s", instance.name, dst_node.name, msg)
-      fin_resu = False
+      feedback_fn("Finalizing export on %s" % dst_node.name)
+      result = self.rpc.call_finalize_export(dst_node.name, instance,
+                                             snap_disks)
+      fin_resu = True
+      msg = result.fail_msg
+      if msg:
+        self.LogWarning("Could not finalize export for instance %s"
+                        " on node %s: %s", instance.name, dst_node.name, msg)
+        fin_resu = False
+
+    finally:
+      if activate_disks:
+        feedback_fn("Deactivating disks for %s" % instance.name)
+        _ShutdownInstanceDisks(self, instance)
 
     nodelist = self.cfg.GetNodeList()
     nodelist.remove(dst_node.name)
@@ -7775,7 +8277,7 @@ class LURemoveExport(NoHooksLU):
                   " Domain Name.")
 
 
-class TagsLU(NoHooksLU):
+class TagsLU(NoHooksLU): # pylint: disable-msg=W0223
   """Generic tags LU.
 
   This is an abstract class which is the parent of all the other tags LUs.
@@ -7788,14 +8290,14 @@ class TagsLU(NoHooksLU):
       name = self.cfg.ExpandNodeName(self.op.name)
       if name is None:
         raise errors.OpPrereqError("Invalid node name (%s)" %
-                                   (self.op.name,))
+                                   (self.op.name,), errors.ECODE_NOENT)
       self.op.name = name
       self.needed_locks[locking.LEVEL_NODE] = name
     elif self.op.kind == constants.TAG_INSTANCE:
       name = self.cfg.ExpandInstanceName(self.op.name)
       if name is None:
         raise errors.OpPrereqError("Invalid instance name (%s)" %
-                                   (self.op.name,))
+                                   (self.op.name,), errors.ECODE_NOENT)
       self.op.name = name
       self.needed_locks[locking.LEVEL_INSTANCE] = name
 
@@ -7811,7 +8313,7 @@ class TagsLU(NoHooksLU):
       self.target = self.cfg.GetInstanceInfo(self.op.name)
     else:
       raise errors.OpPrereqError("Wrong tag type requested (%s)" %
-                                 str(self.op.kind))
+                                 str(self.op.kind), errors.ECODE_INVAL)
 
 
 class LUGetTags(TagsLU):
@@ -7848,7 +8350,7 @@ class LUSearchTags(NoHooksLU):
       self.re = re.compile(self.op.pattern)
     except re.error, err:
       raise errors.OpPrereqError("Invalid search pattern '%s': %s" %
-                                 (self.op.pattern, err))
+                                 (self.op.pattern, err), errors.ECODE_INVAL)
 
   def Exec(self, feedback_fn):
     """Returns the tag list.
@@ -7894,12 +8396,7 @@ class LUAddTags(TagsLU):
         self.target.AddTag(tag)
     except errors.TagError, err:
       raise errors.OpExecError("Error while setting tag: %s" % str(err))
-    try:
-      self.cfg.Update(self.target)
-    except errors.ConfigurationError:
-      raise errors.OpRetryError("There has been a modification to the"
-                                " config file and the operation has been"
-                                " aborted. Please retry.")
+    self.cfg.Update(self.target, feedback_fn)
 
 
 class LUDelTags(TagsLU):
@@ -7925,7 +8422,7 @@ class LUDelTags(TagsLU):
       diff_names = ["'%s'" % tag for tag in diff_tags]
       diff_names.sort()
       raise errors.OpPrereqError("Tag(s) %s not found" %
-                                 (",".join(diff_names)))
+                                 (",".join(diff_names)), errors.ECODE_NOENT)
 
   def Exec(self, feedback_fn):
     """Remove the tag from the object.
@@ -7933,12 +8430,7 @@ class LUDelTags(TagsLU):
     """
     for tag in self.op.tags:
       self.target.RemoveTag(tag)
-    try:
-      self.cfg.Update(self.target)
-    except errors.ConfigurationError:
-      raise errors.OpRetryError("There has been a modification to the"
-                                " config file and the operation has been"
-                                " aborted. Please retry.")
+    self.cfg.Update(self.target, feedback_fn)
 
 
 class LUTestDelay(NoHooksLU):
@@ -7996,6 +8488,8 @@ class IAllocator(object):
       easy usage
 
   """
+  # pylint: disable-msg=R0902
+  # lots of instance attributes
   _ALLO_KEYS = [
     "mem_size", "disks", "disk_template",
     "os", "tags", "nics", "vcpus", "hypervisor",
@@ -8214,10 +8708,12 @@ class IAllocator(object):
                                    " IAllocator" % self.name)
 
     if instance.disk_template not in constants.DTS_NET_MIRROR:
-      raise errors.OpPrereqError("Can't relocate non-mirrored instances")
+      raise errors.OpPrereqError("Can't relocate non-mirrored instances",
+                                 errors.ECODE_INVAL)
 
     if len(instance.secondary_nodes) != 1:
-      raise errors.OpPrereqError("Instance has not exactly one secondary node")
+      raise errors.OpPrereqError("Instance has not exactly one secondary node",
+                                 errors.ECODE_STATE)
 
     self.required_nodes = 1
     disk_sizes = [{'size': disk.size} for disk in instance.disks]
@@ -8305,51 +8801,55 @@ class LUTestAllocator(NoHooksLU):
                    "os", "tags", "nics", "vcpus"]:
         if not hasattr(self.op, attr):
           raise errors.OpPrereqError("Missing attribute '%s' on opcode input" %
-                                     attr)
+                                     attr, errors.ECODE_INVAL)
       iname = self.cfg.ExpandInstanceName(self.op.name)
       if iname is not None:
         raise errors.OpPrereqError("Instance '%s' already in the cluster" %
-                                   iname)
+                                   iname, errors.ECODE_EXISTS)
       if not isinstance(self.op.nics, list):
-        raise errors.OpPrereqError("Invalid parameter 'nics'")
+        raise errors.OpPrereqError("Invalid parameter 'nics'",
+                                   errors.ECODE_INVAL)
       for row in self.op.nics:
         if (not isinstance(row, dict) or
             "mac" not in row or
             "ip" not in row or
             "bridge" not in row):
-          raise errors.OpPrereqError("Invalid contents of the"
-                                     " 'nics' parameter")
+          raise errors.OpPrereqError("Invalid contents of the 'nics'"
+                                     " parameter", errors.ECODE_INVAL)
       if not isinstance(self.op.disks, list):
-        raise errors.OpPrereqError("Invalid parameter 'disks'")
+        raise errors.OpPrereqError("Invalid parameter 'disks'",
+                                   errors.ECODE_INVAL)
       for row in self.op.disks:
         if (not isinstance(row, dict) or
             "size" not in row or
             not isinstance(row["size"], int) or
             "mode" not in row or
             row["mode"] not in ['r', 'w']):
-          raise errors.OpPrereqError("Invalid contents of the"
-                                     " 'disks' parameter")
+          raise errors.OpPrereqError("Invalid contents of the 'disks'"
+                                     " parameter", errors.ECODE_INVAL)
       if not hasattr(self.op, "hypervisor") or self.op.hypervisor is None:
         self.op.hypervisor = self.cfg.GetHypervisorType()
     elif self.op.mode == constants.IALLOCATOR_MODE_RELOC:
       if not hasattr(self.op, "name"):
-        raise errors.OpPrereqError("Missing attribute 'name' on opcode input")
+        raise errors.OpPrereqError("Missing attribute 'name' on opcode input",
+                                   errors.ECODE_INVAL)
       fname = self.cfg.ExpandInstanceName(self.op.name)
       if fname is None:
         raise errors.OpPrereqError("Instance '%s' not found for relocation" %
-                                   self.op.name)
+                                   self.op.name, errors.ECODE_NOENT)
       self.op.name = fname
       self.relocate_from = self.cfg.GetInstanceInfo(fname).secondary_nodes
     else:
       raise errors.OpPrereqError("Invalid test allocator mode '%s'" %
-                                 self.op.mode)
+                                 self.op.mode, errors.ECODE_INVAL)
 
     if self.op.direction == constants.IALLOCATOR_DIR_OUT:
       if not hasattr(self.op, "allocator") or self.op.allocator is None:
-        raise errors.OpPrereqError("Missing allocator name")
+        raise errors.OpPrereqError("Missing allocator name",
+                                   errors.ECODE_INVAL)
     elif self.op.direction != constants.IALLOCATOR_DIR_IN:
       raise errors.OpPrereqError("Wrong allocator test '%s'" %
-                                 self.op.direction)
+                                 self.op.direction, errors.ECODE_INVAL)
 
   def Exec(self, feedback_fn):
     """Run the allocator test.