LGTM

On 10/19/2015 03:57 PM, 'Klaus Aehlig' via ganeti-devel wrote:
In this way, we can support the use case that only some resources
are considered important and dynamic load data is provided for, whereas
the others are ignored completely.

Signed-off-by: Klaus Aehlig <[email protected]>
---
  man/hbal.rst                      | 8 ++++++++
  src/Ganeti/HTools/Program/Hbal.hs | 3 ++-
  2 files changed, 10 insertions(+), 1 deletion(-)

diff --git a/man/hbal.rst b/man/hbal.rst
index a6747c8..8a7d684 100644
--- a/man/hbal.rst
+++ b/man/hbal.rst
@@ -31,6 +31,7 @@ Algorithm options:
  **[ \--avoid-disk-moves *factor* ]**
  **[ \--no-instance-moves ]**
  **[ -U *util-file* ]**
+**[ \--idle-default ]**
  **[ \--ignore-dynu ]**
  **[ \--ignore-soft-errors ]**
  **[ \--mond *yes|no* ]**
@@ -421,6 +422,13 @@ The options that can be passed to the program are as 
follows:
    metrics and thus the influence of the dynamic utilisation will be
    practically insignificant.
+\--idle-default
+  If given, all dynamic utilisation information not provided explicitly
+  by the ``-U`` option or by the MonDs, if ``--mond`` is given, will be
+  assumed to be 0. Note that without this option the default assumption
+  about utilization will apply for the unspecified resources, which is 1.0,
+  i.e., full load, for every instance.
+
  \--ignore-dynu
    If given, all dynamic utilisation information will be ignored by
    assuming it to be 0. This option will take precedence over any data
diff --git a/src/Ganeti/HTools/Program/Hbal.hs 
b/src/Ganeti/HTools/Program/Hbal.hs
index 21b9e03..e2efa57 100644
--- a/src/Ganeti/HTools/Program/Hbal.hs
+++ b/src/Ganeti/HTools/Program/Hbal.hs
@@ -4,7 +4,7 @@
{- -Copyright (C) 2009, 2010, 2011, 2012, 2013 Google Inc.
+Copyright (C) 2009, 2010, 2011, 2012, 2013, 2015 Google Inc.
  All rights reserved.
Redistribution and use in source and binary forms, with or without
@@ -106,6 +106,7 @@ options = do
      , oInstMoves
      , oIgnoreSoftErrors
      , oDynuFile
+    , oIdleDefault
      , oIgnoreDyn
      , oMonD
      , oMonDDataFile

Reply via email to