ganeti.masterd.iallocator.IAllocator

class documentation

class IAllocator(object):

IAllocator framework.

An IAllocator instance has three sets of attributes:

cfg that is needed to query the cluster
input data (all members of the _KEYS class attribute are required)
four buffer attributes (in|out_data|text), that represent the input (to the external script) in text and data structure format, and the output from it, again in two formats
the result variables from the script (success, info, nodes) for easy usage

Method	`__init__`	Undocumented
Method	`Run`	Run an instance allocator and return the results.
Instance Variable	`cfg`	Undocumented
Instance Variable	`in_data`	Undocumented
Instance Variable	`in_text`	Undocumented
Instance Variable	`info`	Undocumented
Instance Variable	`out_data`	Undocumented
Instance Variable	`out_text`	Undocumented
Instance Variable	`req`	Undocumented
Instance Variable	`result`	Undocumented
Instance Variable	`rpc`	Undocumented
Instance Variable	`success`	Undocumented
Static Method	`_ComputeBasicNodeData`	Compute global node data.
Static Method	`_ComputeInstanceData`	Compute global instance data.
Static Method	`_ComputeInstanceMemory`	Compute memory used by primary instances.
Static Method	`_ComputeNodeGroupData`	Compute node groups data.
Static Method	`_ComputeStorageDataFromSpaceInfo`	Extract storage data from node info.
Static Method	`_ComputeStorageDataFromSpaceInfoByTemplate`	Extract storage data from node info.
Static Method	`_GetAttributeFromHypervisorNodeData`	Extract an attribute from the hypervisor's node information.
Method	`_BuildInputData`	Build input data structures.
Method	`_ComputeClusterData`	Compute the generic allocator input data.
Method	`_ComputeClusterDataNodeInfo`	Prepare and execute node info call.
Method	`_ComputeDynamicNodeData`	Compute global node data.
Method	`_ValidateResult`	Process the allocator results.

def __init__(self, cfg, rpc_runner, req): ¶

Undocumented

def Run(self, name, validate=True, call_fn=None): ¶

Run an instance allocator and return the results.

cfg = ¶

Undocumented

in_data = ¶

Undocumented

in_text = ¶

Undocumented

info = ¶

Undocumented

out_data = ¶

Undocumented

out_text = ¶

Undocumented

req = ¶

Undocumented

result = ¶

Undocumented

rpc = ¶

Undocumented

success = ¶

Undocumented

@staticmethod
def _ComputeBasicNodeData(cfg, node_cfg): ¶

Compute global node data.

Returns
dict	a dict of name: (node dict, node config)

@staticmethod
def _ComputeInstanceData(cfg, cluster_info, i_list): ¶

Compute global instance data.

@staticmethod
def _ComputeInstanceMemory(instance_list, node_instances_info, node_uuid, input_mem_free): ¶

Compute memory used by primary instances.

Returns
tuple (int, int, int)	A tuple of three integers: 1. the sum of memory used by primary instances on the node (including the ones that are currently down), 2. the sum of memory used by primary instances of the node that are up, 3. the amount of memory that is free on the node considering the current usage of the instances.

@staticmethod
def _ComputeNodeGroupData(cluster, ginfo): ¶

Compute node groups data.

@staticmethod
def _ComputeStorageDataFromSpaceInfo(space_info, node_name, has_lvm): ¶

Extract storage data from node info.

Parameters
space_info:see result of the RPC call node info	the storage reporting part of the result of the RPC call node info
node_name:string	the node's name
has_lvm:boolean	whether or not LVM storage information is requested
Returns
4-tuple of integers	tuple of storage info (total_disk, free_disk, total_spindles, free_spindles)

@staticmethod
def _ComputeStorageDataFromSpaceInfoByTemplate(space_info, node_name, disk_template): ¶

Extract storage data from node info.

Parameters
space_info:see result of the RPC call node info	the storage reporting part of the result of the RPC call node info
node_name:string	the node's name
disk_template:string	the disk template to report space for
Returns
4-tuple of integers	tuple of storage info (total_disk, free_disk, total_spindles, free_spindles)

@staticmethod
def _GetAttributeFromHypervisorNodeData(hv_info, node_name, attr): ¶

Extract an attribute from the hypervisor's node information.

This is a helper function to extract data from the hypervisor's information about the node, as part of the result of a node_info query.

Parameters
hv_info:dict of strings	dictionary of node information from the hypervisor
node_name:string	name of the node
attr:string	key of the attribute in the hv_info dictionary
Returns
integer	the value of the attribute
Raises
`errors.OpExecError`	if key not in dictionary or value not integer

def _BuildInputData(self, req): ¶

Build input data structures.

def _ComputeClusterData(self, disk_template=None): ¶

Compute the generic allocator input data.

Parameters
disk_template:list of string	the disk templates of the instances to be allocated

def _ComputeClusterDataNodeInfo(self, disk_templates, node_list, cluster_info, hypervisor_name): ¶

Prepare and execute node info call.

Parameters
disk_templates:list of string	the disk templates of the instances to be allocated
node_list:list of strings	list of nodes' UUIDs
cluster_info:`objects.Cluster`	the cluster's information from the config
hypervisor_name:string	the hypervisor name
Returns
same as the result of the node info RPC call	the result of the node info RPC call

def _ComputeDynamicNodeData(self, node_cfg, node_data, node_iinfo, i_list, node_results, disk_template): ¶

Compute global node data.

Parameters
node_cfg	Undocumented
node_data	Undocumented
node_iinfo	Undocumented
i_list	Undocumented
node_results	the basic node structures as filled from the config
disk_template	Undocumented

def _ValidateResult(self): ¶

Process the allocator results.

This will process and if successful save the result in self.out_data and the other parameters.