New module vmware_deploy_ovf (#30309)

* First pass at vmware_deploy_ovf functionality
* Add OVA file support, re-structure code
* Move some useful functions to module_utils.vmware, and perform a little DRY too
* Better handling of errors during spec validation and import
* Properly calculate the lease progress percentage for all vmdk files
* Make warnings and errors a little better
* Add an allow_duplicates argument, that defaults to true, to allow users to have name based idempotency
* Add fail_on_spec_warnings to cause the module to treat warnings as errors
* Support non-vmdk uploads
* Add ova alias for ovf
* Rename vmdk_post_url to device_upload_url so it does not sound to specific to VMDK files
* Safer handling of * hostname in urls
* Add default Content-Type, remove unused headers var
* Add deploymentOptions and propertyMapping functionalities
* Add basic check_mode support
* Add vmware_deploy_ovf to list of use-argspec-type-path ignores
* Update version_added and fix path for use-argspec-type-path
* Add configurable folder
* Doc changes
pull/40834/head
Matt Martz 7 years ago committed by Abhijeet Kasurde
parent 2e560f1ffc
commit 36c8441c6d

@ -45,6 +45,19 @@ def wait_for_task(task):
time.sleep(15)
def wait_for_vm_ip(content, vm, timeout=300):
facts = dict()
interval = 15
while timeout > 0:
facts = gather_vm_facts(content, vm)
if facts['ipv4'] or facts['ipv6']:
break
time.sleep(interval)
timeout -= interval
return facts
def find_obj(content, vimtype, name, first=True):
container = content.viewManager.CreateContainerView(container=content.rootFolder, recursive=True, type=vimtype)
obj_list = container.view
@ -91,6 +104,18 @@ def find_cluster_by_name_datacenter(datacenter, cluster_name):
return None
def find_object_by_name(content, name, obj_type, folder=None, recurse=True):
if not isinstance(obj_type, list):
obj_type = [obj_type]
objects = get_all_objs(content, obj_type, folder=folder, recurse=recurse)
for obj in objects:
if obj.name == name:
return obj
return None
def find_cluster_by_name(content, cluster_name, datacenter=None):
if datacenter:
@ -98,22 +123,11 @@ def find_cluster_by_name(content, cluster_name, datacenter=None):
else:
folder = content.rootFolder
clusters = get_all_objs(content, [vim.ClusterComputeResource], folder)
for cluster in clusters:
if cluster.name == cluster_name:
return cluster
return None
return find_object_by_name(content, cluster_name, [vim.ClusterComputeResource], folder=folder)
def find_datacenter_by_name(content, datacenter_name):
datacenters = get_all_objs(content, [vim.Datacenter])
for dc in datacenters:
if dc.name == datacenter_name:
return dc
return None
return find_object_by_name(content, datacenter_name, [vim.Datacenter])
def get_parent_datacenter(obj):
@ -132,32 +146,23 @@ def get_parent_datacenter(obj):
def find_datastore_by_name(content, datastore_name):
datastores = get_all_objs(content, [vim.Datastore])
for ds in datastores:
if ds.name == datastore_name:
return ds
return None
return find_object_by_name(content, datastore_name, [vim.Datastore])
def find_dvs_by_name(content, switch_name):
return find_object_by_name(content, switch_name, [vim.DistributedVirtualSwitch])
vmware_distributed_switches = get_all_objs(content, [vim.DistributedVirtualSwitch])
for dvs in vmware_distributed_switches:
if dvs.name == switch_name:
return dvs
return None
def find_hostsystem_by_name(content, hostname):
return find_object_by_name(content, hostname, [vim.HostSystem])
def find_hostsystem_by_name(content, hostname):
def find_resource_pool_by_name(content, resource_pool_name):
return find_object_by_name(content, resource_pool_name, [vim.ResourcePool])
host_system = get_all_objs(content, [vim.HostSystem])
for host in host_system:
if host.name == hostname:
return host
return None
def find_network_by_name(content, network_name):
return find_object_by_name(content, network_name, [vim.Network])
def find_vm_by_id(content, vm_id, vm_id_type="vm_name", datacenter=None, cluster=None, folder=None, match_first=False):
@ -197,11 +202,14 @@ def find_vm_by_id(content, vm_id, vm_id_type="vm_name", datacenter=None, cluster
def find_vm_by_name(content, vm_name, folder=None, recurse=True):
return find_object_by_name(content, vm_name, [vim.VirtualMachine], folder=folder, recurse=recurse)
def find_host_portgroup_by_name(host, portgroup_name):
vms = get_all_objs(content, [vim.VirtualMachine], folder, recurse=recurse)
for vm in vms:
if vm.name == vm_name:
return vm
for portgroup in host.config.network.portgroup:
if portgroup.spec.name == portgroup_name:
return portgroup
return None

@ -0,0 +1,586 @@
#!/usr/bin/python
# -*- coding: utf-8 -*-
# Copyright: (c) 2017, Matt Martz <matt@sivel.net>
#
# GNU General Public License v3.0+ (see COPYING or https://www.gnu.org/licenses/gpl-3.0.txt)
from __future__ import absolute_import, division, print_function
__metaclass__ = type
ANSIBLE_METADATA = {
'metadata_version': '1.1',
'status': ['preview'],
'supported_by': 'community'
}
DOCUMENTATION = '''
author: 'Matt Martz (@sivel)'
short_description: 'Deploys a VMware virtual machine from an OVF or OVA file'
description:
- 'This module can be used to deploy a VMware VM from an OVF or OVA file'
module: vmware_deploy_ovf
notes: []
options:
allow_duplicates:
default: "yes"
description:
- Whether or not to allow duplicate VM names. ESXi allows duplicates, vCenter may not.
type: bool
datacenter:
default: ha-datacenter
description:
- Datacenter to deploy to.
datastore:
default: datastore1
description:
- Datastore to deploy to.
deployment_option:
description:
- The key of the chosen deployment option.
disk_provisioning:
choices:
- flat
- eagerZeroedThick
- monolithicSparse
- twoGbMaxExtentSparse
- twoGbMaxExtentFlat
- thin
- sparse
- thick
- seSparse
- monolithicFlat
default: thin
description:
- Disk provisioning type.
fail_on_spec_warnings:
description:
- Cause the module to treat OVF Import Spec warnings as errors.
default: "no"
type: bool
folder:
description:
- Absolute path of folder to place the virtual machine.
- If not specified, defaults to the value of C(datacenter.vmFolder).
name:
description:
- Name of the VM to work with.
- Virtual machine names in vCenter are not necessarily unique, which may be problematic.
networks:
default:
VM Network: VM Network
description:
- 'C(key: value) mapping of OVF network name, to the vCenter network name.'
ovf:
description:
- 'Path to OVF or OVA file to deploy.'
aliases:
- ova
power_on:
default: true
description:
- 'Whether or not to power on the virtual machine after creation.'
type: bool
properties:
description:
- The assignment of values to the properties found in the OVF as key value pairs.
resource_pool:
default: Resources
description:
- 'Resource Pool to deploy to.'
wait:
default: true
description:
- 'Wait for the host to power on.'
type: bool
wait_for_ip_address:
default: false
description:
- Wait until vCenter detects an IP address for the VM.
- This requires vmware-tools (vmtoolsd) to properly work after creation.
type: bool
requirements:
- pyvmomi
version_added: "2.7"
extends_documentation_fragment: vmware.documentation
'''
EXAMPLES = r'''
- vmware_deploy_ovf:
hostname: esx.example.org
username: root
password: passw0rd
ovf: /path/to/ubuntu-16.04-amd64.ovf
wait_for_ip_address: true
'''
RETURN = r'''
instance:
description: metadata about the new virtualmachine
returned: always
type: dict
sample: None
'''
import io
import os
import sys
import tarfile
import time
import traceback
from threading import Thread
from ansible.module_utils._text import to_native
from ansible.module_utils.basic import AnsibleModule
from ansible.module_utils.six import string_types
from ansible.module_utils.urls import generic_urlparse, open_url, urlparse, urlunparse
from ansible.module_utils.vmware import (HAS_PYVMOMI, connect_to_api, find_datacenter_by_name, find_datastore_by_name,
find_network_by_name, find_resource_pool_by_name, find_vm_by_name, gather_vm_facts,
vmware_argument_spec, wait_for_task, wait_for_vm_ip)
try:
from ansible.module_utils.vmware import vim
from pyVmomi import vmodl
except ImportError:
pass
def path_exists(value):
if not isinstance(value, string_types):
value = str(value)
value = os.path.expanduser(os.path.expandvars(value))
if not os.path.exists(value):
raise ValueError('%s is not a valid path' % value)
return value
class ProgressReader(io.FileIO):
def __init__(self, name, mode='r', closefd=True):
self.bytes_read = 0
io.FileIO.__init__(self, name, mode=mode, closefd=closefd)
def read(self, size=10240):
chunk = io.FileIO.read(self, size)
self.bytes_read += len(chunk)
return chunk
class TarFileProgressReader(tarfile.ExFileObject):
def __init__(self, *args):
self.bytes_read = 0
tarfile.ExFileObject.__init__(self, *args)
def __enter__(self):
return self
def __exit__(self, exc_type, exc_value, traceback):
try:
self.close()
except:
pass
def read(self, size=10240):
chunk = tarfile.ExFileObject.read(self, size)
self.bytes_read += len(chunk)
return chunk
class VMDKUploader(Thread):
def __init__(self, vmdk, url, validate_certs=True, tarinfo=None, create=False):
Thread.__init__(self)
self.vmdk = vmdk
if tarinfo:
self.size = tarinfo.size
else:
self.size = os.stat(vmdk).st_size
self.url = url
self.validate_certs = validate_certs
self.tarinfo = tarinfo
self.f = None
self.e = None
self._create = create
@property
def bytes_read(self):
try:
return self.f.bytes_read
except AttributeError:
return 0
def _request_opts(self):
'''
Requests for vmdk files differ from other file types. Build the request options here to handle that
'''
headers = {
'Content-Length': self.size,
'Content-Type': 'application/octet-stream',
}
if self._create:
# Non-VMDK
method = 'PUT'
headers['Overwrite'] = 't'
else:
# VMDK
method = 'POST'
headers['Content-Type'] = 'application/x-vnd.vmware-streamVmdk'
return {
'method': method,
'headers': headers,
}
def _open_url(self):
open_url(self.url, data=self.f, validate_certs=self.validate_certs, **self._request_opts())
def run(self):
if self.tarinfo:
try:
with TarFileProgressReader(self.vmdk, self.tarinfo) as self.f:
self._open_url()
except Exception:
self.e = sys.exc_info()
else:
try:
with ProgressReader(self.vmdk, 'rb') as self.f:
self._open_url()
except Exception:
self.e = sys.exc_info()
class VMwareDeployOvf:
def __init__(self, module):
self.si = connect_to_api(module)
self.module = module
self.params = module.params
self.datastore = None
self.datacenter = None
self.resource_pool = None
self.network_mappings = []
self.ovf_descriptor = None
self.tar = None
self.lease = None
self.import_spec = None
self.entity = None
def get_objects(self):
self.datastore = find_datastore_by_name(self.si, self.params['datastore'])
if not self.datastore:
self.module.fail_json(msg='%(datastore)s could not be located' % self.params)
self.datacenter = find_datacenter_by_name(self.si, self.params['datacenter'])
if not self.datacenter:
self.module.fail_json(msg='%(datacenter)s could not be located' % self.params)
self.resource_pool = find_resource_pool_by_name(self.si, self.params['resource_pool'])
if not self.resource_pool:
self.module.fail_json(msg='%(resource_pool)s could not be located' % self.params)
for key, value in self.params['networks'].items():
network = find_network_by_name(self.si, value)
if not network:
self.module.fail_json(msg='%(network)s could not be located' % self.params)
network_mapping = vim.OvfManager.NetworkMapping()
network_mapping.name = key
network_mapping.network = network
self.network_mappings.append(network_mapping)
return self.datastore, self.datacenter, self.resource_pool, self.network_mappings
def get_ovf_descriptor(self):
if tarfile.is_tarfile(self.params['ovf']):
self.tar = tarfile.open(self.params['ovf'])
ovf = None
for candidate in self.tar.getmembers():
dummy, ext = os.path.splitext(candidate.name)
if ext.lower() == '.ovf':
ovf = candidate
break
if not ovf:
self.module.fail_json(msg='Could not locate OVF file in %(ovf)s' % self.params)
self.ovf_descriptor = to_native(self.tar.extractfile(ovf).read())
else:
with open(self.params['ovf']) as f:
self.ovf_descriptor = f.read()
return self.ovf_descriptor
def get_lease(self):
datastore, datacenter, resource_pool, network_mappings = self.get_objects()
params = {
'diskProvisioning': self.params['disk_provisioning'],
}
if self.params['name']:
params['entityName'] = self.params['name']
if network_mappings:
params['networkMapping'] = network_mappings
if self.params['deployment_option']:
params['deploymentOption'] = self.params['deployment_option']
if self.params['properties']:
params['propertyMapping'] = []
for key, value in self.params['properties'].items():
property_mapping = vim.KeyValue()
property_mapping.key = key
property_mapping.value = value
params['propertyMapping'].append(property_mapping)
if self.params['folder']:
folder = self.si.searchIndex.FindByInventoryPath(self.params['folder'])
else:
folder = datacenter.vmFolder
spec_params = vim.OvfManager.CreateImportSpecParams(**params)
ovf_descriptor = self.get_ovf_descriptor()
self.import_spec = self.si.ovfManager.CreateImportSpec(
ovf_descriptor,
resource_pool,
datastore,
spec_params
)
errors = [to_native(e.msg) for e in getattr(self.import_spec, 'error', [])]
if self.params['fail_on_spec_warnings']:
errors.extend(
(to_native(w.msg) for w in getattr(self.import_spec, 'warning', []))
)
if errors:
self.module.fail_json(
msg='Failure validating OVF import spec: %s' % '. '.join(errors)
)
for warning in getattr(self.import_spec, 'warning', []):
self.module.warn('Problem validating OVF import spec: %s' % to_native(warning.msg))
if not self.params['allow_duplicates']:
name = self.import_spec.importSpec.configSpec.name
match = find_vm_by_name(self.si, name, folder=folder)
if match:
self.module.exit_json(instance=gather_vm_facts(self.si, match), changed=False)
if self.module.check_mode:
self.module.exit_json(changed=True, instance={'hw_name': name})
try:
self.lease = resource_pool.ImportVApp(
self.import_spec.importSpec,
folder
)
except vmodl.fault.SystemError as e:
self.module.fail_json(
msg='Failed to start import: %s' % to_native(e.msg)
)
while self.lease.state != vim.HttpNfcLease.State.ready:
time.sleep(0.1)
self.entity = self.lease.info.entity
return self.lease, self.import_spec
def _normalize_url(self, url):
'''
The hostname in URLs from vmware may be ``*`` update it accordingly
'''
url_parts = generic_urlparse(urlparse(url))
if url_parts.hostname == '*':
if url_parts.port:
url_parts.netloc = '%s:%d' % (self.params['hostname'], url_parts.port)
else:
url_parts.netloc = self.params['hostname']
return urlunparse(url_parts.as_list())
def upload(self):
ovf_dir = os.path.dirname(self.params['ovf'])
lease, import_spec = self.get_lease()
uploaders = []
for file_item in import_spec.fileItem:
device_upload_url = None
for device_url in lease.info.deviceUrl:
if file_item.deviceId == device_url.importKey:
device_upload_url = self._normalize_url(device_url.url)
break
if not device_upload_url:
lease.HttpNfcLeaseAbort(
vmodl.fault.SystemError(reason='Failed to find deviceUrl for file %s' % file_item.path)
)
self.module.fail_json(
msg='Failed to find deviceUrl for file %s' % file_item.path
)
if self.tar:
vmdk = self.tar
try:
vmdk_tarinfo = self.tar.getmember(file_item.path)
except KeyError:
lease.HttpNfcLeaseAbort(
vmodl.fault.SystemError(reason='Failed to find VMDK file %s in OVA' % file_item.path)
)
self.module.fail_json(
msg='Failed to find VMDK file %s in OVA' % file_item.path
)
else:
vmdk = os.path.join(ovf_dir, file_item.path)
try:
path_exists(vmdk)
except ValueError:
lease.HttpNfcLeaseAbort(
vmodl.fault.SystemError(reason='Failed to find VMDK file at %s' % vmdk)
)
self.module.fail_json(
msg='Failed to find VMDK file at %s' % vmdk
)
vmdk_tarinfo = None
uploaders.append(
VMDKUploader(
vmdk,
device_upload_url,
self.params['validate_certs'],
tarinfo=vmdk_tarinfo,
create=file_item.create
)
)
total_size = sum(u.size for u in uploaders)
total_bytes_read = [0] * len(uploaders)
for i, uploader in enumerate(uploaders):
uploader.start()
while uploader.is_alive():
time.sleep(0.1)
total_bytes_read[i] = uploader.bytes_read
lease.HttpNfcLeaseProgress(int(100.0 * sum(total_bytes_read) / total_size))
if uploader.e:
lease.HttpNfcLeaseAbort(
vmodl.fault.SystemError(reason='%s' % to_native(uploader.e[1]))
)
self.module.fail_json(
msg='%s' % to_native(uploader.e[1]),
exception=''.join(traceback.format_tb(uploader.e[2]))
)
def complete(self):
self.lease.HttpNfcLeaseComplete()
def power_on(self):
facts = {}
if self.params['power_on']:
task = self.entity.PowerOn()
if self.params['wait']:
wait_for_task(task)
if self.params['wait_for_ip_address']:
_facts = wait_for_vm_ip(self.si, self.entity)
if not _facts:
self.module.fail_json(msg='Waiting for IP address timed out')
facts.update(_facts)
if not facts:
facts.update(gather_vm_facts(self.si, self.entity))
return facts
def main():
argument_spec = vmware_argument_spec()
argument_spec.update({
'name': {},
'datastore': {
'default': 'datastore1',
},
'datacenter': {
'default': 'ha-datacenter',
},
'deployment_option': {
'default': None,
},
'folder': {
'default': None,
},
'resource_pool': {
'default': 'Resources',
},
'networks': {
'default': {
'VM Network': 'VM Network',
},
'type': 'dict',
},
'ovf': {
'type': path_exists,
'aliases': ['ova'],
},
'disk_provisioning': {
'choices': [
'flat',
'eagerZeroedThick',
'monolithicSparse',
'twoGbMaxExtentSparse',
'twoGbMaxExtentFlat',
'thin',
'sparse',
'thick',
'seSparse',
'monolithicFlat'
],
'default': 'thin',
},
'power_on': {
'type': 'bool',
'default': True,
},
'properties': {
'type': 'dict',
},
'wait': {
'type': 'bool',
'default': True,
},
'wait_for_ip_address': {
'type': 'bool',
'default': False,
},
'allow_duplicates': {
'type': 'bool',
'default': True,
},
'fail_on_spec_warnings': {
'type': 'bool',
'default': False,
},
})
module = AnsibleModule(
argument_spec=argument_spec,
supports_check_mode=True,
)
if not HAS_PYVMOMI:
module.fail_json(msg='pyvmomi python library not found')
deploy_ovf = VMwareDeployOvf(module)
deploy_ovf.upload()
deploy_ovf.complete()
facts = deploy_ovf.power_on()
module.exit_json(instance=facts, changed=True)
if __name__ == '__main__':
main()

@ -20,6 +20,7 @@ def main():
'lib/ansible/modules/web_infrastructure/ansible_tower/tower_host.py',
'lib/ansible/modules/web_infrastructure/ansible_tower/tower_group.py',
'lib/ansible/modules/web_infrastructure/jenkins_plugin.py',
'lib/ansible/modules/cloud/vmware/vmware_deploy_ovf.py',
# fix uses of expanduser in the following modules and remove them from the following list
'lib/ansible/modules/cloud/rackspace/rax.py',
'lib/ansible/modules/cloud/rackspace/rax_scaling_group.py',

Loading…
Cancel
Save