So far, for computing reboot groups, hroller only considered
drbd-instances. Even though the non-drbd instances can be
migrated to any node, we have to account for capacity. As
suggested in the design document as future plan, this is done
by greedily searching for locations where to place the non-drbd
instances.

Also, two options are added for alternative behavior. One option
is to skip nodes that host non-redundant instances when scheduling
reboots. The other option is to ignore the fact that the cluster
hosts non-redundant instances.

Klaus Aehlig (5):
  Remove trailing whitespace
  hroller: option to skip nodes with non-redundant instances
  Make hroller also plan for non-redundant instances
  hroller: option to ignore non-redundant instances
  Test for hroller taking non-redundant instances into account

 Makefile.am                                |  1 +
 src/Ganeti/HTools/CLI.hs                   | 20 +++++++
 src/Ganeti/HTools/Program/Hroller.hs       | 95 +++++++++++++++++++++++++++---
 test/data/htools/hroller-nonredundant.data | 25 ++++++++
 test/hs/shelltests/htools-hroller.test     | 15 +++++
 5 files changed, 149 insertions(+), 7 deletions(-)
 create mode 100644 test/data/htools/hroller-nonredundant.data

-- 
1.8.2.1

Reply via email to