Revision 708f8301

b/lib/masterd/iallocator.py
761 761
    """
762 762
    request = req.GetRequest(self.cfg)
763 763
    disk_template = None
764
    # FIXME: decide this based on request mode
765 764
    if "disk_template" in request:
766 765
      disk_template = request["disk_template"]
767 766
    self._ComputeClusterData(disk_template=disk_template)
b/man/gnt-instance.rst
992 992
    # gnt-instance batch-create instances.json
993 993
    Submitted jobs 37, 38
994 994

  
995

  
996
Note: If the allocator is used for computing suitable nodes for the
997
instances, it will only take into account disk information for the
998
default disk template. That means, even if other disk templates are
999
specified for the instances, storage space information of these disk
1000
templates will not be considered in the allocation computation.
1001

  
1002

  
995 1003
REMOVE
996 1004
^^^^^^
997 1005

  
......
1921 1929

  
1922 1930
This command moves an instance to another node group. The move is
1923 1931
calculated by an iallocator, either given on the command line or as a
1924
cluster default.
1932
cluster default. Note that the iallocator does only consider disk
1933
information of the default disk template, even if the instances'
1934
disk templates differ from that.
1925 1935

  
1926 1936
If no specific destination groups are specified using ``--to``, all
1927 1937
groups except the one containing the instance are considered.
b/man/gnt-node.rst
123 123
  instance that the node is a secondary for.
124 124
- when neither of the above is done a combination of the two cases is run
125 125

  
126
Note that the iallocator currently only considers disk information of
127
the default disk template, even if the instance's disk templates differ
128
from that.
129

  
126 130
See **ganeti**\(7) for a description of ``--submit`` and other common
127 131
options.
128 132

  

Also available in: Unified diff