{-| Module describing a node.

    All updates are functional (copy-based) and return a new node with
    updated value.
-}

{-

Copyright (C) 2009, 2010, 2011, 2012, 2013 Google Inc.
All rights reserved.

Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are
met:

1. Redistributions of source code must retain the above copyright notice,
this list of conditions and the following disclaimer.

2. Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.

THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS
IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR
CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

-}

module Ganeti.HTools.Node
  ( Node(..)
  , List
  , pCpuEff
  , pCpuEffForth
  -- * Constructor
  , create
  -- ** Finalization after data loading
  , buildPeers
  , setIdx
  , setAlias
  , setOffline
  , setXmem
  , setPri
  , calcFmemOfflineOrForthcoming
  , setSec
  , setMaster
  , setNodeTags
  , setMdsk
  , setMcpu
  , setPolicy
  , setCpuSpeed
  , setMigrationTags
  , setRecvMigrationTags
  , setLocationTags
  -- * Tag maps
  , addTags
  , delTags
  , rejectAddTags
  -- * Diagnostic commands
  , getPolicyHealth
  -- * Instance (re)location
  , removePri
  , removeSec
  , addPri
  , addPriEx
  , addSec
  , addSecEx
  , addSecExEx
  , checkMigration
  -- * Stats
  , availDisk
  , availMem
  , availCpu
  , iMem
  , iDsk
  , conflictingPrimaries
  -- * Generate OpCodes
  , genPowerOnOpCodes
  , genPowerOffOpCodes
  , genAddTagsOpCode
  -- * Formatting
  , defaultFields
  , showHeader
  , showField
  , list
  -- * Misc stuff
  , AssocList
  , AllocElement
  , noSecondary
  , computeGroups
  , mkNodeGraph
  , mkRebootNodeGraph
  , haveExclStorage
  ) where

import Control.Monad (liftM, liftM2)
import Control.Applicative ((<$>), (<*>))
import qualified Data.Foldable as Foldable
import Data.Function (on)
import qualified Data.Graph as Graph
import qualified Data.IntMap as IntMap
import Data.List hiding (group)
import qualified Data.Map as Map
import Data.Ord (comparing)
import qualified Data.Set as Set
import Text.Printf (printf)

import qualified Ganeti.Constants as C
import qualified Ganeti.OpCodes as OpCodes
import Ganeti.Types (OobCommand(..), TagKind(..), mkNonEmpty)
import Ganeti.HTools.Container (Container)
import qualified Ganeti.HTools.Container as Container
import Ganeti.HTools.Instance (Instance)
import qualified Ganeti.HTools.Instance as Instance
import qualified Ganeti.HTools.PeerMap as P

import Ganeti.BasicTypes
import qualified Ganeti.HTools.Types as T

-- * Type declarations

-- | The tag map type.
type TagMap = Map.Map String Int

-- | The node type.
data Node = Node
  { name     :: String    -- ^ The node name
  , alias    :: String    -- ^ The shortened name (for display purposes)
  , tMem     :: Double    -- ^ Total memory (MiB)
  , nMem     :: Int       -- ^ Node memory (MiB)
  , fMem     :: Int       -- ^ Free memory (MiB)
  , fMemForth :: Int      -- ^ Free memory (MiB) including forthcoming
                          --   instances
  , xMem     :: Int       -- ^ Unaccounted memory (MiB)
  , tDsk     :: Double    -- ^ Total disk space (MiB)
  , fDsk     :: Int       -- ^ Free disk space (MiB)
  , fDskForth :: Int      -- ^ Free disk space (MiB) including forthcoming
                          --   instances
  , tCpu     :: Double    -- ^ Total CPU count
  , tCpuSpeed :: Double   -- ^ Relative CPU speed
  , nCpu     :: Int       -- ^ VCPUs used by the node OS
  , uCpu     :: Int       -- ^ Used VCPU count
  , uCpuForth :: Int      -- ^ Used VCPU count including forthcoming instances
  , tSpindles :: Int      -- ^ Node spindles (spindle_count node parameter,
                          -- or actual spindles, see note below)
  , fSpindles :: Int      -- ^ Free spindles (see note below)
  , fSpindlesForth :: Int -- ^ Free spindles (see note below) including
                          --   forthcoming instances
  , pList    :: [T.Idx]   -- ^ List of primary instance indices
  , pListForth :: [T.Idx] -- ^ List of primary instance indices including
                          --   forthcoming instances
  , sList    :: [T.Idx]   -- ^ List of secondary instance indices
  , sListForth :: [T.Idx] -- ^ List of secondary instance indices including
                          --   forthcoming instances
  , idx      :: T.Ndx     -- ^ Internal index for book-keeping
  , peers    :: P.PeerMap -- ^ Pnode to instance mapping
  , failN1   :: Bool      -- ^ Whether the node has failed n1
  , failN1Forth :: Bool   -- ^ Whether the node has failed n1, including
                          --   forthcoming instances
  , rMem     :: Int       -- ^ Maximum memory needed for failover by
                          -- primaries of this node
  , rMemForth :: Int      -- ^ Maximum memory needed for failover by
                          --   primaries of this node, including forthcoming
                          --   instances
  , pMem     :: Double    -- ^ Percent of free memory
  , pMemForth :: Double   -- ^ Percent of free memory including forthcoming
                          --   instances
  , pDsk     :: Double    -- ^ Percent of free disk
  , pDskForth :: Double   -- ^ Percent of free disk including forthcoming
                          --   instances
  , pRem     :: Double    -- ^ Percent of reserved memory
  , pRemForth :: Double   -- ^ Percent of reserved memory including
                          --   forthcoming instances
  , pCpu     :: Double    -- ^ Ratio of virtual to physical CPUs
  , pCpuForth :: Double   -- ^ Ratio of virtual to physical CPUs including
                          --   forthcoming instances
  , mDsk     :: Double    -- ^ Minimum free disk ratio
  , loDsk    :: Int       -- ^ Autocomputed from mDsk low disk
                          -- threshold
  , hiCpu    :: Int       -- ^ Autocomputed from mCpu high cpu
                          -- threshold
  , hiSpindles :: Double  -- ^ Limit auto-computed from policy spindle_ratio
                          -- and the node spindle count (see note below)
  , instSpindles :: Double -- ^ Spindles used by instances (see note below)
  , instSpindlesForth :: Double -- ^ Spindles used by instances (see note
                                --   below) including forthcoming instances
  , offline  :: Bool      -- ^ Whether the node should not be used for
                          -- allocations and skipped from score
                          -- computations
  , isMaster :: Bool      -- ^ Whether the node is the master node
  , nTags    :: [String]  -- ^ The node tags for this node
  , utilPool :: T.DynUtil -- ^ Total utilisation capacity
  , utilLoad :: T.DynUtil -- ^ Sum of instance utilisation
  , utilLoadForth :: T.DynUtil -- ^ Sum of instance utilisation, including
                               --   forthcoming instances
  , pTags    :: TagMap    -- ^ Primary instance exclusion tags and their
                          --   count, including forthcoming instances
  , group    :: T.Gdx     -- ^ The node's group (index)
  , iPolicy  :: T.IPolicy -- ^ The instance policy (of the node's group)
  , exclStorage :: Bool   -- ^ Effective value of exclusive_storage
  , migTags  :: Set.Set String -- ^ migration-relevant tags
  , rmigTags :: Set.Set String -- ^ migration tags able to receive
  , locationTags :: Set.Set String -- ^ common-failure domains the node belongs
                                   -- to
  , locationScore :: Int
  } deriving (Show, Eq)
{- A note on how we handle spindles

With exclusive storage spindles is a resource, so we track the number of
spindles still available (fSpindles). This is the only reliable way, as some
spindles could be used outside of Ganeti. When exclusive storage is off,
spindles are a way to represent disk I/O pressure, and hence we track the amount
used by the instances. We compare it against 'hiSpindles', computed from the
instance policy, to avoid policy violations. In both cases we store the total
spindles in 'tSpindles'.
-}

instance T.Element Node where
  nameOf = name
  idxOf = idx
  setAlias = setAlias
  setIdx = setIdx
  allNames n = [name n, alias n]

-- | Derived parameter: ratio of virutal to physical CPUs, weighted
-- by CPU speed.
pCpuEff :: Node -> Double
pCpuEff n = pCpu n / tCpuSpeed n

-- | Derived parameter: ratio of virutal to physical CPUs, weighted
-- by CPU speed and taking forthcoming instances into account.
pCpuEffForth :: Node -> Double
pCpuEffForth n = pCpuForth n / tCpuSpeed n

-- | A simple name for the int, node association list.
type AssocList = [(T.Ndx, Node)]

-- | A simple name for a node map.
type List = Container.Container Node

-- | A simple name for an allocation element (here just for logistic
-- reasons).
type AllocElement = (List, Instance.Instance, [Node], T.Score)

-- | Constant node index for a non-moveable instance.
noSecondary :: T.Ndx
noSecondary = -1

-- * Helper functions

-- | Add a tag to a tagmap.
addTag :: TagMap -> String -> TagMap
addTag t s = Map.insertWith (+) s 1 t

-- | Add multiple tags.
addTags :: TagMap -> [String] -> TagMap
addTags = foldl' addTag

-- | Adjust or delete a tag from a tagmap.
delTag :: TagMap -> String -> TagMap
delTag t s = Map.update (\v -> if v > 1
                                 then Just (v-1)
                                 else Nothing)
             s t

-- | Remove multiple tags.
delTags :: TagMap -> [String] -> TagMap
delTags = foldl' delTag

-- | Check if we can add a list of tags to a tagmap.
rejectAddTags :: TagMap -> [String] -> Bool
rejectAddTags t = any (`Map.member` t)

-- | Check how many primary instances have conflicting tags. The
-- algorithm to compute this is to sum the count of all tags, then
-- subtract the size of the tag map (since each tag has at least one,
-- non-conflicting instance); this is equivalent to summing the
-- values in the tag map minus one.
conflictingPrimaries :: Node -> Int
conflictingPrimaries (Node { pTags = t }) = Foldable.sum t - Map.size t

-- | Helper function to increment a base value depending on the passed
-- boolean argument.
incIf :: (Num a) => Bool -> a -> a -> a
incIf True  base delta = base + delta
incIf False base _     = base

-- | Helper function to decrement a base value depending on the passed
-- boolean argument.
decIf :: (Num a) => Bool -> a -> a -> a
decIf True  base delta = base - delta
decIf False base _     = base

-- | Is exclusive storage enabled on any node?
haveExclStorage :: List -> Bool
haveExclStorage nl =
  any exclStorage $ Container.elems nl

-- * Initialization functions

-- | Create a new node.
--
-- The index and the peers maps are empty, and will be need to be
-- update later via the 'setIdx' and 'buildPeers' functions.
create :: String -> Double -> Int -> Int
       -> Double -> Int -> Double -> Int -> Bool
       -> Int -> Int -> T.Gdx -> Bool
       -> Node
create name_init mem_t_init mem_n_init mem_f_init
       dsk_t_init dsk_f_init cpu_t_init cpu_n_init offline_init
       spindles_t_init spindles_f_init group_init excl_stor =
  Node { name = name_init
       , alias = name_init
       , tMem = mem_t_init
       , nMem = mem_n_init
       , fMem = mem_f_init
       , fMemForth = mem_f_init
       , tDsk = dsk_t_init
       , fDsk = dsk_f_init
       , fDskForth = dsk_f_init
       , tCpu = cpu_t_init
       , tCpuSpeed = 1
       , nCpu = cpu_n_init
       , uCpu = cpu_n_init
       , uCpuForth = cpu_n_init
       , tSpindles = spindles_t_init
       , fSpindles = spindles_f_init
       , fSpindlesForth = spindles_f_init
       , pList = []
       , pListForth = []
       , sList = []
       , sListForth = []
       , failN1 = True
       , failN1Forth = True
       , idx = -1
       , peers = P.empty
       , rMem = 0
       , rMemForth = 0
       , pMem = fromIntegral mem_f_init / mem_t_init
       , pMemForth = fromIntegral mem_f_init / mem_t_init
       , pDsk = if excl_stor
                then computePDsk spindles_f_init $ fromIntegral spindles_t_init
                else computePDsk dsk_f_init dsk_t_init
       , pDskForth =
           if excl_stor
             then computePDsk spindles_f_init $ fromIntegral spindles_t_init
             else computePDsk dsk_f_init dsk_t_init
       , pRem = 0
       , pRemForth = 0
       , pCpu = fromIntegral cpu_n_init / cpu_t_init
       , pCpuForth = fromIntegral cpu_n_init / cpu_t_init
       , offline = offline_init
       , isMaster = False
       , nTags = []
       , xMem = 0
       , mDsk = T.defReservedDiskRatio
       , loDsk = mDskToloDsk T.defReservedDiskRatio dsk_t_init
       , hiCpu = mCpuTohiCpu (T.iPolicyVcpuRatio T.defIPolicy) cpu_t_init
       , hiSpindles = computeHiSpindles (T.iPolicySpindleRatio T.defIPolicy)
                      spindles_t_init
       , instSpindles = 0
       , instSpindlesForth = 0
       , utilPool = T.baseUtil
       , utilLoad = T.zeroUtil
       , utilLoadForth = T.zeroUtil
       , pTags = Map.empty
       , group = group_init
       , iPolicy = T.defIPolicy
       , exclStorage = excl_stor
       , migTags = Set.empty
       , rmigTags = Set.empty
       , locationTags = Set.empty
       , locationScore = 0
       }

-- | Conversion formula from mDsk\/tDsk to loDsk.
mDskToloDsk :: Double -> Double -> Int
mDskToloDsk mval = floor . (mval *)

-- | Conversion formula from mCpu\/tCpu to hiCpu.
mCpuTohiCpu :: Double -> Double -> Int
mCpuTohiCpu mval = floor . (mval *)

-- | Conversiojn formula from spindles and spindle ratio to hiSpindles.
computeHiSpindles :: Double -> Int -> Double
computeHiSpindles spindle_ratio = (spindle_ratio *) . fromIntegral

-- | Changes the index.
--
-- This is used only during the building of the data structures.
setIdx :: Node -> T.Ndx -> Node
setIdx t i = t {idx = i}

-- | Changes the alias.
--
-- This is used only during the building of the data structures.
setAlias :: Node -> String -> Node
setAlias t s = t { alias = s }

-- | Sets the offline attribute.
setOffline :: Node -> Bool -> Node
setOffline t val = t { offline = val }

-- | Sets the master attribute
setMaster :: Node -> Bool -> Node
setMaster t val = t { isMaster = val }

-- | Sets the node tags attribute
setNodeTags :: Node -> [String] -> Node
setNodeTags t val = t { nTags = val }

-- | Set migration tags
setMigrationTags :: Node -> Set.Set String -> Node
setMigrationTags t val = t { migTags = val }

-- | Set the migration tags a node is able to receive
setRecvMigrationTags :: Node -> Set.Set String -> Node
setRecvMigrationTags t val = t { rmigTags = val }

-- | Set the location tags
setLocationTags :: Node -> Set.Set String -> Node
setLocationTags t val = t { locationTags = val }

-- | Sets the unnaccounted memory.
setXmem :: Node -> Int -> Node
setXmem t val = t { xMem = val }

-- | Sets the max disk usage ratio.
setMdsk :: Node -> Double -> Node
setMdsk t val = t { mDsk = val, loDsk = mDskToloDsk val (tDsk t) }

-- | Sets the max cpu usage ratio. This will update the node's
-- ipolicy, losing sharing (but it should be a seldomly done operation).
setMcpu :: Node -> Double -> Node
setMcpu t val =
  let new_ipol = (iPolicy t) { T.iPolicyVcpuRatio = val }
  in t { hiCpu = mCpuTohiCpu val (tCpu t), iPolicy = new_ipol }

-- | Sets the policy.
setPolicy :: T.IPolicy -> Node -> Node
setPolicy pol node =
  node { iPolicy = pol
       , hiCpu = mCpuTohiCpu (T.iPolicyVcpuRatio pol) (tCpu node)
       , hiSpindles = computeHiSpindles (T.iPolicySpindleRatio pol)
                      (tSpindles node)
       }

-- | Computes the maximum reserved memory for peers from a peer map.
computeMaxRes :: P.PeerMap -> P.Elem
computeMaxRes = P.maxElem

-- | Builds the peer map for a given node.
buildPeers :: Node -> Instance.List -> Node
buildPeers t il =
  let mdata = map
              (\i_idx -> let inst = Container.find i_idx il
                             mem = if Instance.usesSecMem inst
                                      -- TODO Use usesMemory here, or change
                                      --      usesSecMem to return False on
                                      --      forthcoming instances?
                                      && not (Instance.forthcoming inst)
                                     then Instance.mem inst
                                     else 0
                         in (Instance.pNode inst, mem))
              (sList t)
      pmap = P.accumArray (+) mdata
      new_rmem = computeMaxRes pmap
      new_failN1 = fMem t < new_rmem
      new_prem = fromIntegral new_rmem / tMem t
  in t { peers = pmap
       , failN1 = new_failN1
       , rMem = new_rmem
       , pRem = new_prem

       -- TODO Set failN1Forth, rMemForth, pRemForth and peersForth.
       --      Calculate it from an mdata_forth here that doesn't have the
       --      `not (Instance.forthcoming inst)` filter.
       }

-- | Calculate the new spindle usage
calcSpindleUse ::
                  Bool -- Action: True = adding instance, False = removing it
               -> Node -> Instance.Instance -> Double
calcSpindleUse _ (Node {exclStorage = True}) _ = 0.0
calcSpindleUse act n@(Node {exclStorage = False}) i =
  f (Instance.usesLocalStorage i) (instSpindles n)
    (fromIntegral $ Instance.spindleUse i)
    where
      f :: Bool -> Double -> Double -> Double -- avoid monomorphism restriction
      f = if act then incIf else decIf

-- | Calculate the new spindle usage including forthcoming instances.
calcSpindleUseForth :: Bool -- Action: True = adding instance, False = removing
                    -> Node -> Instance.Instance -> Double
calcSpindleUseForth _ (Node {exclStorage = True}) _ = 0.0
calcSpindleUseForth act n@(Node {exclStorage = False}) i =
  f (Instance.usesLocalStorage i) (instSpindlesForth n)
    (fromIntegral $ Instance.spindleUse i)
    where
      f :: Bool -> Double -> Double -> Double -- avoid monomorphism restriction
      f = if act then incIf else decIf

-- | Calculate the new number of free spindles
calcNewFreeSpindles ::
                       Bool -- Action: True = adding instance, False = removing
                    -> Node -> Instance.Instance -> Int
calcNewFreeSpindles _ (Node {exclStorage = False}) _ = 0
calcNewFreeSpindles act n@(Node {exclStorage = True}) i =
  case Instance.getTotalSpindles i of
    Nothing -> if act
               then -1 -- Force a spindle error, so the instance don't go here
               else fSpindles n -- No change, as we aren't sure
    Just s -> (if act then (-) else (+)) (fSpindles n) s

-- | Calculate the new number of free spindles including forthcoming instances
calcNewFreeSpindlesForth :: Bool -- Action: True = adding instance,
                                 --         False = removing
                         -> Node -> Instance.Instance -> Int
calcNewFreeSpindlesForth _ (Node {exclStorage = False}) _ = 0
calcNewFreeSpindlesForth act n@(Node {exclStorage = True}) i =
  case Instance.getTotalSpindles i of
    Nothing -> if act
               then -1 -- Force a spindle error, so the instance don't go here
               else fSpindlesForth n -- No change, as we aren't sure
    Just s -> (if act then (-) else (+)) (fSpindlesForth n) s


calcFmemOfflineOrForthcoming :: Node -> Container Instance -> Int
calcFmemOfflineOrForthcoming node allInstances =
  let nodeInstances = map (`Container.find` allInstances) (pList node)
  in sum . map Instance.mem
         . filter (not . Instance.usesMemory)
         $ nodeInstances

-- | Assigns an instance to a node as primary and update the used VCPU
-- count, utilisation data and tags map.
setPri :: Node -> Instance.Instance -> Node
setPri t inst
  -- Real instance, update real fields and forthcoming fields.
  | not (Instance.forthcoming inst) =
      updateForthcomingFields $
        t { pList = Instance.idx inst:pList t
          , uCpu = new_count
          , pCpu = fromIntegral new_count / tCpu t
          , utilLoad = utilLoad t `T.addUtil` Instance.util inst
          , instSpindles = calcSpindleUse True t inst
          , locationScore = locationScore t + Instance.locationScore inst
          }

  -- Forthcoming instance, update forthcoming fields only.
  | otherwise = updateForthcomingOnlyFields $ updateForthcomingFields t

  where
    new_count = Instance.applyIfOnline inst (+ Instance.vcpus inst) (uCpu t)
    new_count_forth = Instance.applyIfOnline inst (+ Instance.vcpus inst)
                                             (uCpuForth t)

    uses_disk = Instance.usesLocalStorage inst

    -- Updates the *Forth fields that include real and forthcoming instances.
    updateForthcomingFields node =

      let new_fMemForth = decIf (not $ Instance.usesMemory inst)
                                (fMemForth node)
                                (Instance.mem inst)

          new_pMemForth = fromIntegral new_fMemForth / tMem node

      in node
           { pTags = addTags (pTags node) (Instance.exclTags inst)

           , pListForth = Instance.idx inst:pListForth node
           , uCpuForth = new_count_forth
           , pCpuForth = fromIntegral new_count_forth / tCpu node
           , utilLoadForth = utilLoadForth node `T.addUtil` Instance.util inst

           , fMemForth = new_fMemForth
           , pMemForth = new_pMemForth

           -- TODO Should this be in updateForthcomingOnlyFields?
           , instSpindlesForth = calcSpindleUseForth True node inst

           -- TODO Set failN1Forth, rMemForth, pRemForth
           }

    -- This updates the fields that we do not want to update if the instance
    -- is real (not forthcoming), in contrast to `updateForthcomingFields`
    -- which deals with the fields that have to be updated in either case.
    updateForthcomingOnlyFields node =

      let new_fDskForth = decIf uses_disk
                                (fDskForth node)
                                (Instance.dsk inst)

          new_free_sp_forth = calcNewFreeSpindlesForth True node inst
          new_pDskForth = computeNewPDsk node new_free_sp_forth new_fDskForth

      in node
           { fDskForth = new_fDskForth
           , pDskForth = new_pDskForth
           , fSpindlesForth = new_free_sp_forth
           }


-- | Assigns an instance to a node as secondary and updates disk utilisation.
setSec :: Node -> Instance.Instance -> Node
setSec t inst
  -- Real instance, update real fields and forthcoming fields.
  | not (Instance.forthcoming inst) =
      updateForthcomingFields $
        t { sList = Instance.idx inst:sList t
          , utilLoad = old_load { T.dskWeight = T.dskWeight old_load +
                                  T.dskWeight (Instance.util inst) }
          , instSpindles = calcSpindleUse True t inst
          }

  -- Forthcoming instance, update forthcoming fields only.
  | otherwise = updateForthcomingOnlyFields $ updateForthcomingFields t

  where
    old_load = utilLoad t
    uses_disk = Instance.usesLocalStorage inst

    -- Updates the *Forth fields that include real and forthcoming instances.
    updateForthcomingFields node =

      let old_load_forth = utilLoadForth node
      in node
           { sListForth = Instance.idx inst:sListForth node
           , utilLoadForth = old_load_forth
                               { T.dskWeight = T.dskWeight old_load_forth +
                                               T.dskWeight (Instance.util inst)
                               }

           -- TODO Should this be in updateForthcomingOnlyFields?
           , instSpindlesForth = calcSpindleUseForth True node inst

           -- TODO Set failN1Forth, rMemForth, pRemForth and peersForth
           }

    updateForthcomingOnlyFields node =

      let new_fDskForth = decIf uses_disk
                                (fDskForth node)
                                (Instance.dsk inst)
          new_free_sp_forth = calcNewFreeSpindlesForth True node inst
          new_pDskForth = computeNewPDsk node new_free_sp_forth new_fDskForth
      in node
           { fDskForth = new_fDskForth
           , pDskForth = new_pDskForth
           , fSpindlesForth = new_free_sp_forth
           }


-- | Computes the new 'pDsk' value, handling nodes without local disk
-- storage (we consider all their disk unused).
computePDsk :: Int -> Double -> Double
computePDsk _    0     = 1
computePDsk free total = fromIntegral free / total

-- | Computes the new 'pDsk' value, handling the exclusive storage state.
computeNewPDsk :: Node -> Int -> Int -> Double
computeNewPDsk node new_free_sp new_free_dsk =
  if exclStorage node
  then computePDsk new_free_sp . fromIntegral $ tSpindles node
  else computePDsk new_free_dsk $ tDsk node

-- * Diagnostic functions

-- | For a node diagnose whether it conforms with all policies. The type
-- is chosen to represent that of a no-op node operation.
getPolicyHealth :: Node -> T.OpResult ()
getPolicyHealth n =
  case () of
    _ | instSpindles n > hiSpindles n -> Bad T.FailDisk
      | pCpu n > T.iPolicyVcpuRatio (iPolicy n) -> Bad T.FailCPU
      | otherwise -> Ok ()

-- * Update functions

-- | Set the CPU speed
setCpuSpeed :: Node -> Double -> Node
setCpuSpeed n f = n { tCpuSpeed = f }

-- | Removes a primary instance.
removePri :: Node -> Instance.Instance -> Node
removePri t inst =
  let iname = Instance.idx inst
      forthcoming = Instance.forthcoming inst
      i_online = Instance.notOffline inst
      uses_disk = Instance.usesLocalStorage inst

      updateForthcomingFields n =
        let
            new_plist_forth = delete iname (pListForth n)
            new_mem_forth = fMemForth n + Instance.mem inst
            new_dsk_forth = incIf uses_disk (fDskForth n) (Instance.dsk inst)
            new_free_sp_forth = calcNewFreeSpindlesForth False n inst
            new_inst_sp_forth = calcSpindleUseForth False n inst
            new_mp_forth = fromIntegral new_mem_forth / tMem n
            new_dp_forth = computeNewPDsk n new_free_sp_forth new_dsk_forth
            new_ucpu_forth = decIf i_online (uCpuForth n) (Instance.vcpus inst)
            new_rcpu_forth = fromIntegral new_ucpu_forth / tCpu n
            new_load_forth = utilLoadForth n `T.subUtil` Instance.util inst

        in n { pTags = delTags (pTags t) (Instance.exclTags inst)

             , pListForth = new_plist_forth
             , fMemForth = new_mem_forth
             , fDskForth = new_dsk_forth
             , pMemForth = new_mp_forth
             , pDskForth = new_dp_forth
             , uCpuForth = new_ucpu_forth
             , pCpuForth = new_rcpu_forth
             , utilLoadForth = new_load_forth
             , instSpindlesForth = new_inst_sp_forth
             , fSpindlesForth = new_free_sp_forth

             -- TODO Set failN1Forth, rMemForth, pRemForth
             }

  in if forthcoming
       then updateForthcomingFields t
       else let
                new_plist = delete iname (pList t)
                new_mem = incIf (Instance.usesMemory inst) (fMem t)
                                (Instance.mem inst)
                new_dsk = incIf uses_disk (fDsk t) (Instance.dsk inst)
                new_free_sp = calcNewFreeSpindles False t inst
                new_inst_sp = calcSpindleUse False t inst
                new_mp = fromIntegral new_mem / tMem t
                new_dp = computeNewPDsk t new_free_sp new_dsk
                new_failn1 = new_mem <= rMem t
                new_ucpu = decIf i_online (uCpu t) (Instance.vcpus inst)
                new_rcpu = fromIntegral new_ucpu / tCpu t
                new_load = utilLoad t `T.subUtil` Instance.util inst

            in updateForthcomingFields $
                 t { pList = new_plist, fMem = new_mem, fDsk = new_dsk
                   , failN1 = new_failn1, pMem = new_mp, pDsk = new_dp
                   , uCpu = new_ucpu, pCpu = new_rcpu, utilLoad = new_load
                   , instSpindles = new_inst_sp, fSpindles = new_free_sp
                   , locationScore = locationScore t
                                     - Instance.locationScore inst
                   }

-- | Removes a secondary instance.
removeSec :: Node -> Instance.Instance -> Node
removeSec t inst =
  let iname = Instance.idx inst
      forthcoming = Instance.forthcoming inst
      uses_disk = Instance.usesLocalStorage inst
      cur_dsk = fDsk t
      pnode = Instance.pNode inst

      updateForthcomingFields n =
        let
            new_slist_forth = delete iname (sListForth n)
            new_dsk_forth = incIf uses_disk (fDskForth n) (Instance.dsk inst)
            new_free_sp_forth = calcNewFreeSpindlesForth False n inst
            new_inst_sp_forth = calcSpindleUseForth False n inst
            new_dp_forth = computeNewPDsk n new_free_sp_forth new_dsk_forth
            old_load_forth = utilLoadForth n
            new_load_forth = old_load_forth
                               { T.dskWeight = T.dskWeight old_load_forth -
                                               T.dskWeight (Instance.util inst)
                               }
        in n { sListForth = new_slist_forth
             , fDskForth = new_dsk_forth
             , pDskForth = new_dp_forth
             , utilLoadForth = new_load_forth
             , instSpindlesForth = new_inst_sp_forth
             , fSpindlesForth = new_free_sp_forth

             -- TODO Set failN1Forth, rMemForth, pRemForth
             }

  in if forthcoming
       then updateForthcomingFields t
       else let
                new_slist = delete iname (sList t)
                new_dsk = incIf uses_disk cur_dsk (Instance.dsk inst)
                new_free_sp = calcNewFreeSpindles False t inst
                new_inst_sp = calcSpindleUse False t inst
                old_peers = peers t
                old_peem = P.find pnode old_peers
                new_peem = decIf (Instance.usesSecMem inst) old_peem
                                 (Instance.mem inst)
                new_peers = if new_peem > 0
                              then P.add pnode new_peem old_peers
                              else P.remove pnode old_peers
                old_rmem = rMem t
                new_rmem = if old_peem < old_rmem
                             then old_rmem
                             else computeMaxRes new_peers
                new_prem = fromIntegral new_rmem / tMem t
                new_failn1 = fMem t <= new_rmem
                new_dp = computeNewPDsk t new_free_sp new_dsk
                old_load = utilLoad t
                new_load = old_load
                  { T.dskWeight = T.dskWeight old_load
                                    - T.dskWeight (Instance.util inst)
                  }
            in updateForthcomingFields $
                 t { sList = new_slist, fDsk = new_dsk, peers = new_peers
                   , failN1 = new_failn1, rMem = new_rmem, pDsk = new_dp
                   , pRem = new_prem, utilLoad = new_load
                   , instSpindles = new_inst_sp, fSpindles = new_free_sp
                   }

-- | Adds a primary instance (basic version).
addPri :: Node -> Instance.Instance -> T.OpResult Node
addPri = addPriEx False

-- | Adds a primary instance (extended version).
addPriEx :: Bool               -- ^ Whether to override the N+1 and
                               -- other /soft/ checks, useful if we
                               -- come from a worse status (e.g. offline).
                               -- If this is True, forthcoming instances
                               -- may exceed available Node resources.
         -> Node               -- ^ The target node
         -> Instance.Instance  -- ^ The instance to add
         -> T.OpResult Node    -- ^ The result of the operation,
                               -- either the new version of the node
                               -- or a failure mode
addPriEx force t inst =
  let iname = Instance.idx inst
      forthcoming = Instance.forthcoming inst
      i_online = Instance.notOffline inst
      uses_disk = Instance.usesLocalStorage inst
      l_cpu = T.iPolicyVcpuRatio $ iPolicy t
      old_tags = pTags t
      strict = not force

      inst_tags = Instance.exclTags inst

      new_mem_forth = fMemForth t - Instance.mem inst
      new_mp_forth = fromIntegral new_mem_forth / tMem t
      new_dsk_forth = decIf uses_disk (fDskForth t) (Instance.dsk inst)
      new_free_sp_forth = calcNewFreeSpindlesForth True t inst
      new_inst_sp_forth = calcSpindleUseForth True t inst
      new_ucpu_forth = incIf i_online (uCpuForth t) (Instance.vcpus inst)
      new_pcpu_forth = fromIntegral new_ucpu_forth / tCpu t
      new_dp_forth = computeNewPDsk t new_free_sp_forth new_dsk_forth
      new_load_forth = utilLoadForth t `T.addUtil` Instance.util inst
      new_plist_forth = iname:pListForth t

      updateForthcomingFields n =
        n { pTags = addTags old_tags inst_tags

          , pListForth = new_plist_forth
          , fMemForth = new_mem_forth
          , fDskForth = new_dsk_forth
          , pMemForth = new_mp_forth
          , pDskForth = new_dp_forth
          , uCpuForth = new_ucpu_forth
          , pCpuForth = new_pcpu_forth
          , utilLoadForth = new_load_forth
          , instSpindlesForth = new_inst_sp_forth
          , fSpindlesForth = new_free_sp_forth

          -- TODO Set failN1Forth, rMemForth, pRemForth
          }

      checkForthcomingViolation
        | new_mem_forth <= 0                            = Bad T.FailMem
        | uses_disk && new_dsk_forth <= 0               = Bad T.FailDisk
        | uses_disk && new_dsk_forth < loDsk t          = Bad T.FailDisk
        | uses_disk && exclStorage t
                    && new_free_sp_forth < 0            = Bad T.FailSpindles
        | uses_disk && new_inst_sp_forth > hiSpindles t = Bad T.FailDisk
        -- TODO Check failN1 including forthcoming instances
        | l_cpu >= 0 && l_cpu < new_pcpu_forth          = Bad T.FailCPU
        | otherwise                                     = Ok ()

  in
    if forthcoming
      then case strict of
             True | Bad err <- checkForthcomingViolation -> Bad err

             _ -> Ok $ updateForthcomingFields t

      else let
               new_mem = decIf (Instance.usesMemory inst) (fMem t)
                               (Instance.mem inst)
               new_dsk = decIf uses_disk (fDsk t) (Instance.dsk inst)
               new_free_sp = calcNewFreeSpindles True t inst
               new_inst_sp = calcSpindleUse True t inst
               new_failn1 = new_mem <= rMem t
               new_ucpu = incIf i_online (uCpu t) (Instance.vcpus inst)
               new_pcpu = fromIntegral new_ucpu / tCpu t
               new_dp = computeNewPDsk t new_free_sp new_dsk
               new_load = utilLoad t `T.addUtil` Instance.util inst

               new_plist = iname:pList t
               new_mp = fromIntegral new_mem / tMem t
      in case () of
        _ | new_mem <= 0 -> Bad T.FailMem
          | uses_disk && new_dsk <= 0 -> Bad T.FailDisk
          | strict && uses_disk && new_dsk < loDsk t -> Bad T.FailDisk
          | uses_disk && exclStorage t && new_free_sp < 0 -> Bad T.FailSpindles
          | strict && uses_disk && new_inst_sp > hiSpindles t -> Bad T.FailDisk
          | strict && new_failn1 && not (failN1 t) -> Bad T.FailMem
          | strict && l_cpu >= 0 && l_cpu < new_pcpu -> Bad T.FailCPU
          | strict && rejectAddTags old_tags inst_tags -> Bad T.FailTags

          -- When strict also check forthcoming limits, but after normal checks
          | strict, Bad err <- checkForthcomingViolation -> Bad err

          | otherwise ->
              Ok . updateForthcomingFields $
                t { pList = new_plist
                  , fMem = new_mem
                  , fDsk = new_dsk
                  , failN1 = new_failn1
                  , pMem = new_mp
                  , pDsk = new_dp
                  , uCpu = new_ucpu
                  , pCpu = new_pcpu
                  , utilLoad = new_load
                  , instSpindles = new_inst_sp
                  , fSpindles = new_free_sp
                  , locationScore = locationScore t
                                    + Instance.locationScore inst
                  }

-- | Adds a secondary instance (basic version).
addSec :: Node -> Instance.Instance -> T.Ndx -> T.OpResult Node
addSec = addSecEx False

-- | Adds a secondary instance (extended version).
addSecEx :: Bool -> Node -> Instance.Instance -> T.Ndx -> T.OpResult Node
addSecEx = addSecExEx False

-- | Adds a secondary instance (doubly extended version). The first parameter
-- tells `addSecExEx` to ignore disks completly. There is only one legitimate
-- use case for this, and this is failing over a DRBD instance where the primary
-- node is offline (and hence will become the secondary afterwards).
addSecExEx :: Bool
           -> Bool -> Node -> Instance.Instance -> T.Ndx -> T.OpResult Node
addSecExEx ignore_disks force t inst pdx =
  let iname = Instance.idx inst
      forthcoming = Instance.forthcoming inst
      old_peers = peers t
      strict = not force

      secondary_needed_mem = if Instance.usesSecMem inst
                               then Instance.mem inst
                               else 0
      new_peem = P.find pdx old_peers + secondary_needed_mem
      new_peers = P.add pdx new_peem old_peers

      old_mem_forth = fMemForth t
      new_dsk_forth = fDskForth t - Instance.dsk inst
      new_free_sp_forth = calcNewFreeSpindlesForth True t inst
      new_inst_sp_forth = calcSpindleUseForth True t inst
      new_dp_forth = computeNewPDsk t new_free_sp_forth new_dsk_forth
      old_load_forth = utilLoadForth t
      new_load_forth = old_load_forth
                         { T.dskWeight = T.dskWeight old_load_forth +
                                         T.dskWeight (Instance.util inst)
                         }
      new_slist_forth = iname:sListForth t

      updateForthcomingFields n =
        n { sListForth = new_slist_forth
          , fDskForth = new_dsk_forth
          , pDskForth = new_dp_forth
          , utilLoadForth = new_load_forth
          , instSpindlesForth = new_inst_sp_forth
          , fSpindlesForth = new_free_sp_forth

          -- TODO Set failN1Forth, rMemForth, pRemForth
          }

      checkForthcomingViolation
        | not (Instance.hasSecondary inst)       = Bad T.FailDisk
        | new_dsk_forth <= 0                     = Bad T.FailDisk
        | new_dsk_forth < loDsk t                = Bad T.FailDisk
        | exclStorage t && new_free_sp_forth < 0 = Bad T.FailSpindles
        | new_inst_sp_forth > hiSpindles t       = Bad T.FailDisk
        | secondary_needed_mem >= old_mem_forth  = Bad T.FailMem
        -- TODO Check failN1 including forthcoming instances
        | otherwise                              = Ok ()

  in if forthcoming
      then case strict of
             True | Bad err <- checkForthcomingViolation -> Bad err

             _ -> Ok $ updateForthcomingFields t
      else let
               old_mem = fMem t
               new_dsk = fDsk t - Instance.dsk inst
               new_free_sp = calcNewFreeSpindles True t inst
               new_inst_sp = calcSpindleUse True t inst
               new_rmem = max (rMem t) new_peem
               new_prem = fromIntegral new_rmem / tMem t
               new_failn1 = old_mem <= new_rmem
               new_dp = computeNewPDsk t new_free_sp new_dsk
               old_load = utilLoad t
               new_load = old_load
                            { T.dskWeight = T.dskWeight old_load +
                                            T.dskWeight (Instance.util inst)
                            }
               new_slist = iname:sList t
      in case () of
        _ | not (Instance.hasSecondary inst) -> Bad T.FailDisk
          | not ignore_disks && new_dsk <= 0 -> Bad T.FailDisk
          | strict && new_dsk < loDsk t -> Bad T.FailDisk
          | exclStorage t && new_free_sp < 0 -> Bad T.FailSpindles
          | strict && new_inst_sp > hiSpindles t -> Bad T.FailDisk
          | strict && secondary_needed_mem >= old_mem -> Bad T.FailMem
          | strict && new_failn1 && not (failN1 t) -> Bad T.FailMem

          -- When strict also check forthcoming limits, but after normal checks
          | strict, Bad err <- checkForthcomingViolation -> Bad err

          | otherwise ->
              Ok . updateForthcomingFields $
                t { sList = new_slist, fDsk = new_dsk
                  , peers = new_peers, failN1 = new_failn1
                  , rMem = new_rmem, pDsk = new_dp
                  , pRem = new_prem, utilLoad = new_load
                  , instSpindles = new_inst_sp
                  , fSpindles = new_free_sp
                  }


-- | Predicate on whether migration is supported between two nodes.
checkMigration :: Node -> Node -> T.OpResult ()
checkMigration nsrc ntarget =
  if migTags nsrc `Set.isSubsetOf` rmigTags ntarget
    then Ok ()
    else Bad T.FailMig

-- * Stats functions

-- | Computes the amount of available disk on a given node.
availDisk :: Node -> Int
availDisk t =
  let _f = fDsk t -- TODO Shall we use fDiskForth here?
      _l = loDsk t
  in if _f < _l
       then 0
       else _f - _l

-- | Computes the amount of used disk on a given node.
iDsk :: Node -> Int
iDsk t = truncate (tDsk t) - fDsk t

-- | Computes the amount of available memory on a given node.
availMem :: Node -> Int
availMem t =
  let _f = fMem t
      _l = rMem t
  in if _f < _l
       then 0
       else _f - _l

-- | Computes the amount of available memory on a given node.
availCpu :: Node -> Int
availCpu t =
  let _u = uCpu t
      _l = hiCpu t
  in if _l >= _u
       then _l - _u
       else 0

-- | The memory used by instances on a given node.
iMem :: Node -> Int
iMem t = truncate (tMem t) - nMem t - xMem t - fMem t

-- * Node graph functions
-- These functions do the transformations needed so that nodes can be
-- represented as a graph connected by the instances that are replicated
-- on them.

-- * Making of a Graph from a node/instance list

-- | Transform an instance into a list of edges on the node graph
instanceToEdges :: Instance.Instance -> [Graph.Edge]
instanceToEdges i
  | Instance.hasSecondary i = [(pnode,snode), (snode,pnode)]
  | otherwise = []
    where pnode = Instance.pNode i
          snode = Instance.sNode i

-- | Transform the list of instances into list of destination edges
instancesToEdges :: Instance.List -> [Graph.Edge]
instancesToEdges = concatMap instanceToEdges . Container.elems

-- | Transform the list of nodes into vertices bounds.
-- Returns Nothing is the list is empty.
nodesToBounds :: List -> Maybe Graph.Bounds
nodesToBounds nl = liftM2 (,) nmin nmax
    where nmin = fmap (fst . fst) (IntMap.minViewWithKey nl)
          nmax = fmap (fst . fst) (IntMap.maxViewWithKey nl)

-- | The clique of the primary nodes of the instances with a given secondary.
-- Return the full graph of those nodes that are primary node of at least one
-- instance that has the given node as secondary.
nodeToSharedSecondaryEdge :: Instance.List -> Node -> [Graph.Edge]
nodeToSharedSecondaryEdge il n = (,) <$> primaries <*> primaries
  where primaries = map (Instance.pNode . flip Container.find il) $ sList n


-- | Predicate of an edge having both vertices in a set of nodes.
filterValid :: List -> [Graph.Edge] -> [Graph.Edge]
filterValid nl  =  filter $ \(x,y) -> IntMap.member x nl && IntMap.member y nl

-- | Transform a Node + Instance list into a NodeGraph type.
-- Returns Nothing if the node list is empty.
mkNodeGraph :: List -> Instance.List -> Maybe Graph.Graph
mkNodeGraph nl il =
  liftM (`Graph.buildG` (filterValid nl . instancesToEdges $ il))
  (nodesToBounds nl)

-- | Transform a Nodes + Instances into a NodeGraph with all reboot exclusions.
-- This includes edges between nodes that are the primary nodes of instances
-- that have the same secondary node. Nodes not in the node list will not be
-- part of the graph, but they are still considered for the edges arising from
-- two instances having the same secondary node.
-- Return Nothing if the node list is empty.
mkRebootNodeGraph :: List -> List -> Instance.List -> Maybe Graph.Graph
mkRebootNodeGraph allnodes nl il =
  liftM (`Graph.buildG` filterValid nl edges) (nodesToBounds nl)
  where
    edges = instancesToEdges il `union`
            (Container.elems allnodes >>= nodeToSharedSecondaryEdge il) 

-- * Display functions

-- | Return a field for a given node.
showField :: Node   -- ^ Node which we're querying
          -> String -- ^ Field name
          -> String -- ^ Field value as string
showField t field =
  case field of
    "idx"  -> printf "%4d" $ idx t
    "name" -> alias t
    "fqdn" -> name t
    "status" -> case () of
                  _ | offline t -> "-"
                    | failN1 t -> "*"
                    | otherwise -> " "
    "tmem" -> printf "%5.0f" $ tMem t
    "nmem" -> printf "%5d" $ nMem t
    "xmem" -> printf "%5d" $ xMem t
    "fmem" -> printf "%5d" $ fMem t
    "imem" -> printf "%5d" $ iMem t
    "rmem" -> printf "%5d" $ rMem t
    "amem" -> printf "%5d" $ fMem t - rMem t
    "tdsk" -> printf "%5.0f" $ tDsk t / 1024
    "fdsk" -> printf "%5d" $ fDsk t `div` 1024
    "tcpu" -> printf "%4.0f" $ tCpu t
    "ucpu" -> printf "%4d" $ uCpu t
    "pcnt" -> printf "%3d" $ length (pList t)
    "scnt" -> printf "%3d" $ length (sList t)
    "plist" -> show $ pList t
    "slist" -> show $ sList t
    "pfmem" -> printf "%6.4f" $ pMem t
    "pfdsk" -> printf "%6.4f" $ pDsk t
    "rcpu"  -> printf "%5.2f" $ pCpu t
    "cload" -> printf "%5.3f" uC
    "mload" -> printf "%5.3f" uM
    "dload" -> printf "%5.3f" uD
    "nload" -> printf "%5.3f" uN
    "ptags" -> intercalate "," . map (uncurry (printf "%s=%d")) .
               Map.toList $ pTags t
    "peermap" -> show $ peers t
    "spindle_count" -> show $ tSpindles t
    "hi_spindles" -> show $ hiSpindles t
    "inst_spindles" -> show $ instSpindles t
    _ -> T.unknownField
  where
    T.DynUtil { T.cpuWeight = uC, T.memWeight = uM,
                T.dskWeight = uD, T.netWeight = uN } = utilLoad t

-- | Returns the header and numeric propery of a field.
showHeader :: String -> (String, Bool)
showHeader field =
  case field of
    "idx" -> ("Index", True)
    "name" -> ("Name", False)
    "fqdn" -> ("Name", False)
    "status" -> ("F", False)
    "tmem" -> ("t_mem", True)
    "nmem" -> ("n_mem", True)
    "xmem" -> ("x_mem", True)
    "fmem" -> ("f_mem", True)
    "imem" -> ("i_mem", True)
    "rmem" -> ("r_mem", True)
    "amem" -> ("a_mem", True)
    "tdsk" -> ("t_dsk", True)
    "fdsk" -> ("f_dsk", True)
    "tcpu" -> ("pcpu", True)
    "ucpu" -> ("vcpu", True)
    "pcnt" -> ("pcnt", True)
    "scnt" -> ("scnt", True)
    "plist" -> ("primaries", True)
    "slist" -> ("secondaries", True)
    "pfmem" -> ("p_fmem", True)
    "pfdsk" -> ("p_fdsk", True)
    "rcpu"  -> ("r_cpu", True)
    "cload" -> ("lCpu", True)
    "mload" -> ("lMem", True)
    "dload" -> ("lDsk", True)
    "nload" -> ("lNet", True)
    "ptags" -> ("PrimaryTags", False)
    "peermap" -> ("PeerMap", False)
    "spindle_count" -> ("NodeSpindles", True)
    "hi_spindles" -> ("MaxSpindles", True)
    "inst_spindles" -> ("InstSpindles", True)
    -- TODO: add node fields (group.uuid, group)
    _ -> (T.unknownField, False)

-- | String converter for the node list functionality.
list :: [String] -> Node -> [String]
list fields t = map (showField t) fields

-- | Generate OpCode for setting a node's offline status
genOpSetOffline :: (Monad m) => Node -> Bool -> m OpCodes.OpCode
genOpSetOffline node offlineStatus = do
  nodeName <- mkNonEmpty (name node)
  return OpCodes.OpNodeSetParams
           { OpCodes.opNodeName = nodeName
           , OpCodes.opNodeUuid = Nothing
           , OpCodes.opForce = False
           , OpCodes.opHvState = Nothing
           , OpCodes.opDiskState = Nothing
           , OpCodes.opMasterCandidate = Nothing
           , OpCodes.opOffline = Just offlineStatus
           , OpCodes.opDrained = Nothing
           , OpCodes.opAutoPromote = False
           , OpCodes.opMasterCapable = Nothing
           , OpCodes.opVmCapable = Nothing
           , OpCodes.opSecondaryIp = Nothing
           , OpCodes.opgenericNdParams = Nothing
           , OpCodes.opPowered = Nothing
           }

-- | Generate OpCode for applying a OobCommand to the given nodes
genOobCommand :: (Monad m) => [Node] -> OobCommand -> m OpCodes.OpCode
genOobCommand nodes command = do
  names <- mapM (mkNonEmpty . name) nodes
  return OpCodes.OpOobCommand
    { OpCodes.opNodeNames = names
    , OpCodes.opNodeUuids = Nothing
    , OpCodes.opOobCommand = command
    , OpCodes.opOobTimeout = C.oobTimeout
    , OpCodes.opIgnoreStatus = False
    , OpCodes.opPowerDelay = C.oobPowerDelay
    }

-- | Generate OpCode for powering on a list of nodes
genPowerOnOpCodes :: (Monad m) => [Node] -> m [OpCodes.OpCode]
genPowerOnOpCodes nodes = do
  opSetParams <- mapM (`genOpSetOffline` False) nodes
  oobCommand <- genOobCommand nodes OobPowerOn
  return $ opSetParams ++ [oobCommand]

-- | Generate OpCodes for powering off a list of nodes
genPowerOffOpCodes :: (Monad m) => [Node] -> m [OpCodes.OpCode]
genPowerOffOpCodes nodes = do
  opSetParams <- mapM (`genOpSetOffline` True) nodes
  oobCommand <- genOobCommand nodes OobPowerOff
  return $ opSetParams ++ [oobCommand]

-- | Generate OpCodes for adding tags to a node
genAddTagsOpCode :: Node -> [String] -> OpCodes.OpCode
genAddTagsOpCode node tags = OpCodes.OpTagsSet
                               { OpCodes.opKind = TagKindNode
                               , OpCodes.opTagsList = tags
                               , OpCodes.opTagsGetName = Just $ name node
                               }

-- | Constant holding the fields we're displaying by default.
defaultFields :: [String]
defaultFields =
  [ "status", "name", "tmem", "nmem", "imem", "xmem", "fmem"
  , "rmem", "tdsk", "fdsk", "tcpu", "ucpu", "pcnt", "scnt"
  , "pfmem", "pfdsk", "rcpu"
  , "cload", "mload", "dload", "nload" ]

{-# ANN computeGroups "HLint: ignore Use alternative" #-}
-- | Split a list of nodes into a list of (node group UUID, list of
-- associated nodes).
computeGroups :: [Node] -> [(T.Gdx, [Node])]
computeGroups nodes =
  let nodes' = sortBy (comparing group) nodes
      nodes'' = groupBy ((==) `on` group) nodes'
  -- use of head here is OK, since groupBy returns non-empty lists; if
  -- you remove groupBy, also remove use of head
  in map (\nl -> (group (head nl), nl)) nodes''