heat-translator/translator/hot/tosca/tosca_compute.py
Julio Ruano 3f98139a7e Allow for dynamic class loading of target types
Tosca types need to be mapped to corresponding target translation
types (i.e. HOT). This commit allows for the target types to be
dynamically loaded from a directory. It allows for more flexibility
by pulling types from a known directory instead of defining each
individiual type statically in the code. For example, with this
commit you no longer need a separate import for each target type.

Also, this commit adds the notion of a global configuration object
that can be shared throughout the translator. The configuration
defines the location directory for custom defined target types.
This configuration can be set by the user in the corresponding
translator/conf/translator.conf file. In the future, additional
values can be added to this configuration and the code can be
extended to support them, but only the required values were
implemented here.

Change-Id: If7b8da12eef5b8ed8a2e11b1f412203d4ed59c5a
Implements: blueprint dynamic-tosca-to-hot-map
2015-09-28 12:09:02 -05:00

212 lines
9.5 KiB
Python
Executable File

#
# Licensed under the Apache License, Version 2.0 (the "License"); you may
# not use this file except in compliance with the License. You may obtain
# a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
# WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
# License for the specific language governing permissions and limitations
# under the License.
import logging
from toscaparser.utils.validateutils import TOSCAVersionProperty
import translator.common.utils
from translator.hot.syntax.hot_resource import HotResource
log = logging.getLogger('tosca')
# Name used to dynamically load appropriate map class.
TARGET_CLASS_NAME = 'ToscaCompute'
# A design issue to be resolved is how to translate the generic TOSCA server
# properties to OpenStack flavors and images. At the Atlanta design summit,
# there was discussion on using Glance to store metadata and Graffiti to
# describe artifacts. We will follow these projects to see if they can be
# leveraged for this TOSCA translation.
# For development purpose at this time, we temporarily hardcode a list of
# flavors and images here
FLAVORS = {'m1.xlarge': {'mem_size': 16384, 'disk_size': 160, 'num_cpus': 8},
'm1.large': {'mem_size': 8192, 'disk_size': 80, 'num_cpus': 4},
'm1.medium': {'mem_size': 4096, 'disk_size': 40, 'num_cpus': 2},
'm1.small': {'mem_size': 2048, 'disk_size': 20, 'num_cpus': 1},
'm1.tiny': {'mem_size': 512, 'disk_size': 1, 'num_cpus': 1},
'm1.micro': {'mem_size': 128, 'disk_size': 0, 'num_cpus': 1},
'm1.nano': {'mem_size': 64, 'disk_size': 0, 'num_cpus': 1}}
IMAGES = {'ubuntu-software-config-os-init': {'architecture': 'x86_64',
'type': 'Linux',
'distribution': 'Ubuntu',
'version': '14.04'},
'ubuntu-12.04-software-config-os-init': {'architecture': 'x86_64',
'type': 'Linux',
'distribution': 'Ubuntu',
'version': '12.04'},
'fedora-amd64-heat-config': {'architecture': 'x86_64',
'type': 'Linux',
'distribution': 'Fedora',
'version': '18.0'},
'F18-x86_64-cfntools': {'architecture': 'x86_64',
'type': 'Linux',
'distribution': 'Fedora',
'version': '19'},
'Fedora-x86_64-20-20131211.1-sda': {'architecture': 'x86_64',
'type': 'Linux',
'distribution': 'Fedora',
'version': '20'},
'cirros-0.3.1-x86_64-uec': {'architecture': 'x86_64',
'type': 'Linux',
'distribution': 'CirrOS',
'version': '0.3.1'},
'cirros-0.3.2-x86_64-uec': {'architecture': 'x86_64',
'type': 'Linux',
'distribution': 'CirrOS',
'version': '0.3.2'},
'rhel-6.5-test-image': {'architecture': 'x86_64',
'type': 'Linux',
'distribution': 'RHEL',
'version': '6.5'}}
class ToscaCompute(HotResource):
'''Translate TOSCA node type tosca.nodes.Compute.'''
toscatype = 'tosca.nodes.Compute'
def __init__(self, nodetemplate):
super(ToscaCompute, self).__init__(nodetemplate,
type='OS::Nova::Server')
# List with associated hot port resources with this server
self.assoc_port_resources = []
pass
def handle_properties(self):
self.properties = self.translate_compute_flavor_and_image(
self.nodetemplate.get_capability('host'),
self.nodetemplate.get_capability('os'))
self.properties['user_data_format'] = 'SOFTWARE_CONFIG'
# TODO(anyone): handle user key
# hardcoded here for testing
self.properties['key_name'] = 'userkey'
# To be reorganized later based on new development in Glance and Graffiti
def translate_compute_flavor_and_image(self,
host_capability,
os_capability):
hot_properties = {}
host_cap_props = {}
os_cap_props = {}
image = None
flavor = None
if host_capability:
for prop in host_capability.get_properties_objects():
host_cap_props[prop.name] = prop.value
flavor = self._best_flavor(host_cap_props)
if os_capability:
for prop in os_capability.get_properties_objects():
os_cap_props[prop.name] = prop.value
image = self._best_image(os_cap_props)
hot_properties['flavor'] = flavor
hot_properties['image'] = image
# TODO(anyone): consider adding the flavor or image as a template
# parameter if no match is found.
return hot_properties
def _best_flavor(self, properties):
# start with all flavors
match_all = FLAVORS.keys()
# TODO(anyone): Handle the case where the value contains something like
# get_input instead of a value.
# flavors that fit the CPU count
cpu = properties.get('num_cpus')
match_cpu = self._match_flavors(match_all, FLAVORS, 'num_cpus', cpu)
# flavors that fit the mem size
mem = properties.get('mem_size')
if mem:
mem = translator.common.utils.MemoryUnit.convert_unit_size_to_num(
mem, 'MB')
match_cpu_mem = self._match_flavors(match_cpu, FLAVORS,
'mem_size', mem)
# flavors that fit the disk size
disk = properties.get('disk_size')
if disk:
disk = translator.common.utils.MemoryUnit.\
convert_unit_size_to_num(disk, 'GB')
match_cpu_mem_disk = self._match_flavors(match_cpu_mem, FLAVORS,
'disk_size', disk)
# if multiple match, pick the flavor with the least memory
# the selection can be based on other heuristic, e.g. pick one with the
# least total resource
if len(match_cpu_mem_disk) > 1:
return self._least_flavor(match_cpu_mem_disk, FLAVORS, 'mem_size')
elif len(match_cpu_mem_disk) == 1:
return match_cpu_mem_disk[0]
else:
return None
def _best_image(self, properties):
match_all = IMAGES.keys()
architecture = properties.get('architecture')
match_arch = self._match_images(match_all, IMAGES,
'architecture', architecture)
type = properties.get('type')
match_type = self._match_images(match_arch, IMAGES, 'type', type)
distribution = properties.get('distribution')
match_distribution = self._match_images(match_type, IMAGES,
'distribution',
distribution)
version = properties.get('version')
version = TOSCAVersionProperty(version).get_version()
match_version = self._match_images(match_distribution, IMAGES,
'version', version)
if len(match_version):
return list(match_version)[0]
def _match_flavors(self, this_list, this_dict, attr, size):
'''Return from this list all flavors matching the attribute size.'''
if not size:
return list(this_list)
matching_flavors = []
for flavor in this_list:
if isinstance(size, int):
if this_dict[flavor][attr] >= size:
matching_flavors.append(flavor)
return matching_flavors
def _least_flavor(self, this_list, this_dict, attr):
'''Return from this list the flavor with the smallest attr.'''
least_flavor = this_list[0]
for flavor in this_list:
if this_dict[flavor][attr] < this_dict[least_flavor][attr]:
least_flavor = flavor
return least_flavor
def _match_images(self, this_list, this_dict, attr, prop):
if not prop:
return this_list
matching_images = []
for image in this_list:
if this_dict[image][attr].lower() == str(prop).lower():
matching_images.append(image)
return matching_images
def get_hot_attribute(self, attribute, args):
attr = {}
# Convert from a TOSCA attribute for a nodetemplate to a HOT
# attribute for the matching resource. Unless there is additional
# runtime support, this should be a one to one mapping.
# Note: We treat private and public IP addresses equally, but
# this will change in the future when TOSCA starts to support
# multiple private/public IP addresses.
if attribute == 'private_address' or \
attribute == 'public_address':
attr['get_attr'] = [self.name, 'networks', 'private', 0]
return attr